Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development and evaluation of luminescent nanoparticles for immunolabeling and imaging via microscopy… Tran, Michael 2020

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2021_may_tran_michael.pdf [ 15.59MB ]
Metadata
JSON: 24-1.0395312.json
JSON-LD: 24-1.0395312-ld.json
RDF/XML (Pretty): 24-1.0395312-rdf.xml
RDF/JSON: 24-1.0395312-rdf.json
Turtle: 24-1.0395312-turtle.txt
N-Triples: 24-1.0395312-rdf-ntriples.txt
Original Record: 24-1.0395312-source.json
Full Text
24-1.0395312-fulltext.txt
Citation
24-1.0395312.ris

Full Text

DEVELOPMENT AND EVALUATION OF LUMINESCENT NANOPARTICLES FOR IMMUNOLABELING AND IMAGING VIA MICROSCOPY  AND SMARTPHONE-BASED DEVICES  by  Michael Tran  B.Sc., McMaster University, 2015  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Chemistry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   December 2020  © Michael Tran, 2020            ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled: Development and Evaluation of Luminescent Nanoparticles for Immunolabeling and Imaging via Microscopy and Smartphone-Based Devices  Submitted by Michael Tran in partial fulfillment of the requirements for the degree of  Doctor of Philosophy in Chemistry  Examining Committee: W. Russ Algar, Chemistry, UBC Supervisor  Dan Bizzotto, Chemistry, UBC Supervisory Committee Member  Jayachandran Kizhakkedathu, Chemistry, UBC Supervisory Committee Member Mark MacLachlan, Chemistry, UBC University Examiner Hongshen Ma, Mechanical Engineering, UBC University Examiner  Additional Supervisory Committee Members: Roman Krems, Chemistry, UBC Supervisory Committee Member     iii  Abstract  Fluorescent nanomaterials are of interest for applications in bioanalysis and imaging due to their potential for high brightness, robust photostability, diverse surface functionalization, and special size-dependent properties. Cellular immunolabeling with antigen-targeting biomolecules (e.g. antibodies, aptamers) is one such application, and has numerous applications in biomedical research and for laboratory-based or point-of-care molecular diagnostics for health care. For example, immunolabeling is a commonly used technique for the diagnostic screening of cancerous cells and relies on the detection of antigens that are specific to malignant cells. However, integrating fluorescent nanomaterials within molecular diagnostic assays can be challenging— optimization to the physicochemical properties of a nanomaterial and its surface chemistry is necessary. When optimized, fluorescent nanoparticles have the ability to address key challenges in bioanalysis and imaging, including enhanced sensitivity and enabling the use of mass-produced consumer technologies, such as smartphones, as a platform for point-of-care diagnostics.   This thesis presents key contributions towards the development of fluorescent nanoparticles for applications in bioanalysis and imaging, and particularly immunofluorescent labeling of cells. These materials include quantum dots (QDs), single-chain polymer nanoparticles (SCPNs), polymer dots (Pdots), and composite supra-nanoparticles that comprise iron oxide nanoparticles/QDs and silica nanoparticle/QDs. Breast cancer cells were labeled and imaged using various antibody conjugates of these nanoparticles. With several of the materials, dextran is demonstrated to be an ideal surface coating. Benefits include improved colloidal stability, reduced non-specific binding, and utility as a biochemical handle for the assembly of tetrameric antibody complexes (TACs) for specific cellular immunolabeling. Despite several advantages, TACs had not been used with fluorescent nanoparticles prior to this thesis. In addition, the exceptional brightness of the supra-nanoparticles enabled their utilization for smartphone-based imaging of immunolabeled cells. This imaging is done with a device manufactured by 3-D printing, which, in conjugation with the brightness of the materials, avoids the typical trade-off between optimal sensitivity and portability, simplicity, and low cost. Overall, this thesis enriches the science of cellular immunolabeling and imaging via novel materials, novel immunoconjugation methods, and novel devices.  iv  Lay Summary  This thesis investigates the use of fluorescent (i.e. light emitting) nanometer-sized particles for visualizing specific proteins on the surfaces of cells. Methods are developed and evaluated with various types of nanoparticles, including those composed of ‘soft’ fluorescent polymers, ‘hard’ inorganic materials, and assemblies of smaller nanoparticles. Across these materials, several methods for the attachment of antibodies—which target specific cell surface proteins—were investigated. Fluorescent nanoparticles with attached antibodies were used to selectively label cancer cells for imaging via research grade microscope or a smartphone-based imaging platform. These novel materials, methods, and technologies have prospective applications as new and improved tools for biomedical research and for simple, portable, and cost-effective health care-related diagnostics (e.g. cancer detection). v  Preface  The research presented in this thesis was done under the supervision of Prof. Russ Algar, who contributed to the conception, design, and analysis of the research and experiments. These contributions were consistent with his supervisory role as principal investigator and are not specially noted. Ethics approval was not required for the research presented within this thesis. My contributions to the research in this thesis and the role of colleagues and collaborators are clarified in the paragraphs below.  The introductory section on quantum dots (Section 1.4.2) is an adaptation of a published book chapter: Wegner, K.D.; Tran, M.V.; Massey, M.; Algar, W.R. (2017) “Quantum Dots in the Analysis of Food Safety and Quality” in Sensing Techniques for Food Safety and Quality Control, Royal Society of Chemistry, pp 17–60. Adapted with permission from the Royal Society of Chemistry. Copyright 2017 the Royal Society of Chemistry). I drafted the original book chapter text adapted for this thesis, with final editing by Prof. Russ Algar.   Chapter 2 is an adaptation of published work: Rees#, K.; Tran#, M.V.; Massey, M.; Kim, H.; Krause, K.D.; Algar, W.R., Dextran-Functionalized Semiconductor Quantum Dot Bioconjugates for Bioanalysis and Imaging. Bioconjugate Chem. 2020, 31, 861–874 (# denotes co-first authors). Adapted with permission from The American Chemical Society. Copyright 2020 American Chemical Society. Kelly Rees and I designed, completed, and analyzed the colourimetric tests for the dextran ligands. I designed, completed, and analyzed cellular immunolabeling studies. Kelly Rees and Melissa Massey designed, synthesized, and characterized the dextran ligands, and contributed physical and photophysical characterization data for the dextran-coated QDs. Melissa Massey contributed immunoassay and protein-based non-specific binding experiments. Kelly Rees and I designed, completed and analyzed the cell-based non-specific binding experiments. Hyungki Kim contributed cellular microinjection results. I designed, completed, and analyzed the pH sensing application. All authors wrote text for the manuscript that was consistent with their experimental contributions, with final editing by Prof. Russ Algar.  vi  Chapter 3 is an adaptation of published work: Bajj, D.N.F.; Tran, M.V.; Tsai, H.-Y.; Kim, H.; Paisley, N.R.; Algar, W.R.; Hudson, Z.M., Fluorescent Heterotelechelic Single-Chain Polymer Nanoparticles: Synthesis, Spectroscopy, and Cellular Imaging. ACS Appl. Nano Mater. 2019, 2, 898–909. Adapted with permission from The American Chemical Society. Copyright 2019 American Chemical Society. This research was done in collaboration with Prof. Zachary Hudson’s research group. Daniel Bajj synthesized the SCPNs. Hsin-Yun Tsai and I designed, completed, and analyzed the cellular immunolabeling and cellular viability experiments. Hsin-Yun Tsai, Hyungki Kim, and I collaborated on the photophysical characterization of the SCPNs. All authors wrote text for the manuscript that was consistent with their experimental contributions, with final editing by Prof. Zachary Hudson and Prof. Russ Algar.   Chapter 4 is based on research done in collaboration with Prof. Zachary Hudson’s research group. I designed, completed, and analyzed the cellular immunolabeling experiments. Chris Tonge (Hudson Research Group) synthesized the HMAT-ODA polymer. Colleagues in the Algar Research Group, Rupsa Gupta and Ghinwa Darwish, contributed photophysical characterization data. Rupsa Gupta prepared the antibody-conjugated HMAT-ODA Pdots.  Chapter 5 is an adaptation of published work: Lix, K.; Tran, M.V.; Massey, M.; Rees, K.; Sauvé, E.R.; Hudson, Z.M.; Algar, W.R., Dextran Functionalization of Semiconducting Polymer Dots and Conjugation with Tetrameric Antibody Complexes for Bioanalysis and Imaging. ACS Appl. Bio Mater. 2020, 3, 432–440. Adapted with permission from The American Chemical Society. Copyright 2020 American Chemical Society. I designed, completed, and analyzed the cellular immunolabeling. Kelsi Lix contributed the photophysical characterization and prepared the Pdot immunoconjugates. All authors wrote text for the manuscript that was consistent with their experimental contributions, with final editing by Kelsi Lix and Prof. Russ Algar.  Chapter 6 is an adaptation of published work: Darwish, G.H.; Asselin, J.; Tran, M.V.; Gupta, R.; Kim, H.; Boudreau, D.; Algar, W.R., Fully Self-Assembled Silica Nanoparticle-Semiconductor Quantum Dot Supra-Nanoparticles and Immunoconjugates for Enhanced Cellular Imaging by Microscopy and Smartphone Camera. ACS Appl. Mater. Interfaces 2020, 12, 33530–33540. Adapted with permission from The American Chemical Society. Copyright 2020 American vii  Chemical Society. Supra-nanoparticles were prepared by Ghinwa Darwish and Jérémie Asselin, who also contributed the physical and photophysical characterization data. Jérémie Asselin and I prepared the dextran-coated supra-NPs, including the synthesis of the modified dextran and the coating of the supra-nanoparticles. I designed and Jérémie Asselin and I completed the cellular immunolabeling experiments, both via fluorescence microscopy and via smartphone-based imaging. I designed and manufactured the smartphone-based imaging platform. Ghinwa Darwish and I analyzed the images of the immunolabeled cells. All authors wrote text for the manuscript that was consistent with their experimental contributions, with final editing by Ghinwa Darwish and Prof. Russ Algar.   Chapter 7 is an adaptation of published work: Tran, M.V., Susumu, K., Medintz, I.L., Algar, W.R., Supraparticle Assemblies of Magnetic Nanoparticles and Quantum Dots for Selective Cell Isolation and Counting on a Smartphone-Based Imaging Platform. Anal. Chem. 2019, 91, 11963–11971. Adapted with permission from The American Chemical Society. Copyright 2019 American Chemical Society. I designed, completed, and analyzed the experiments. Kimihiro Susumu and Igor Medintz provided blue-emitting quantum dots. I wrote the manuscript with input and editing from Prof. Russ Algar.  Chapter 8 is an adaptation of a submitted manuscript: Kim#, H.; Tran#, M.V.; Petryayeva, E.; Solodova, O.; Susumu, K.; Medintz, I.L.; Algar, WR, Affinity Immobilization of Semiconductor Quantum Dots and Metal Nanoparticles on Cellulose Paper Substrates. ACS Appl. Mater. Inter. In review (# denotes co-first authors). Eleonora Petryayeva developed the affinity chemistries. Eleonora Petryayeva, Hyungki Kim, and Olga Solodova prepared paper substrates with immobilized QDs and contributed chemical and photophysical characterization. Hyungki Kim prepared paper substrates with immobilized metal nanoparticles and designed, completed, and analyzed associated experiments. I contributed SEM imaging and designed, completed, and analyzed the immobilization by microcontact printing experiments.   viii  Table of Contents  Abstract ......................................................................................................................................... iii  Lay Summary ............................................................................................................................... iv  Preface .............................................................................................................................................v  Table of Contents ....................................................................................................................... viii  List of Tables ............................................................................................................................. xvii  List of Figures ........................................................................................................................... xviii List of Abbreviations .............................................................................................................. xxxii  Acknowledgements .............................................................................................................. xxxviii   Introduction ................................................................................................................1 1.1 Overview ......................................................................................................................... 1 1.2 Fluorescence ................................................................................................................... 1 1.2.1 The Perrin-Jablonski diagram ..................................................................................... 1  1.2.2 Photon absorption ....................................................................................................... 3 1.2.2.1 Absorption spectrum ........................................................................................... 5 1.2.3 Excited-state processes ............................................................................................... 7  1.2.4 Quantum yield and fluorescence lifetime ................................................................... 9  1.2.5 Photodegradation of a fluorophore ........................................................................... 11 1.2.6 Factors affecting fluorescence .................................................................................. 11 1.2.7 Fluorescence measurements and imaging ................................................................. 12 1.3 Immunohistochemistry ................................................................................................. 14 1.3.1 Chromogenic immunohistochemistry ....................................................................... 15 1.3.2 Immunofluorescence ................................................................................................. 16 1.3.3 Flow cytometry ......................................................................................................... 16 1.4 General overview of nanoparticles ............................................................................... 17 1.4.1 Silica nanoparticles ................................................................................................... 17 1.4.2 Quantum dots ............................................................................................................ 18  1.4.2.1 QDs in bioanalysis and imaging ....................................................................... 20 1.4.2.2 Synthesis and functionalization strategies........................................................ 21 1.4.2.3 Bioconjugation strategies ................................................................................. 22 ix  1.4.2.4 Immunolabeling with QDs ................................................................................ 27 1.4.3 Magnetic nanoparticles ............................................................................................. 30 1.4.4 Metal nanoparticles ................................................................................................... 30 1.4.5 Fluorescent polymeric nanoparticles ........................................................................ 32 1.4.6 Single-chain polymer nanoparticles .......................................................................... 34 1.4.7 Non-specific adsorption on nanoparticles ................................................................. 35 1.5 Point-of-care diagnostics .............................................................................................. 36 1.5.1 Smartphone-based diagnostics .................................................................................. 38 1.5.1.1 Smartphone components ................................................................................... 38 1.5.1.2 Image processing .............................................................................................. 41 1.5.1.3 Utility of smartphones in POC diagnostics ....................................................... 42 1.6 Overview and contributions of this thesis ..................................................................... 44  Dextran-functionalized semiconductor quantum dot-bioconjugates for bioanalysis and imaging ..............................................................................................................48  2.1 Introduction ................................................................................................................... 48  2.2 Results and discussion .................................................................................................. 50 2.2.1 Synthetic strategies ................................................................................................... 50 2.2.2 Ligand characterization ............................................................................................. 53 2.2.3 Characterization of dextran-functionalized QDs ...................................................... 53 2.2.3.1 Surface chemistry.............................................................................................. 54 2.2.3.2 Optical characterization .................................................................................... 56 2.2.3.3 Size characterization ......................................................................................... 56 2.2.3.4 Electrophoretic mobility ................................................................................... 58 2.2.3.5 Colloidal stability .............................................................................................. 61 2.2.4 Assessing non-specific binding................................................................................. 62 2.2.4.1 Protein binding via gel electrophoresis ............................................................. 62 2.2.4.2 Binding to cells ................................................................................................. 64 2.2.5 Microinjection and cell viability ............................................................................... 66 2.2.6 Covalent conjugation and pH sensing....................................................................... 68  2.2.7 Immunolabeling with tetrameric antibody complex conjugates ............................... 70 2.2.8 Discussion ................................................................................................................. 72  x  2.3 Conclusion .................................................................................................................... 74 2.4 Experimental section ..................................................................................................... 74 2.4.1 Materials ................................................................................................................... 75  2.4.2 Instrumentation ......................................................................................................... 77  2.4.3 Dextran ligands and QD functionalization ............................................................... 78 2.4.4 Functional tests ......................................................................................................... 78  2.4.4.1 Anthrone assay with QDs ................................................................................. 78 2.4.4.2 Tests with ConA, BSA, and lysozyme.............................................................. 79 2.4.5 Electrophoretic mobility ........................................................................................... 79 2.4.5.1 Agarose gel electrophoresis .............................................................................. 79 2.4.5.2 Capillary electrophoresis .................................................................................. 79 2.4.6 Cell-based methods ................................................................................................... 80 2.4.6.1 Cell viability...................................................................................................... 80 2.4.6.2 Non-specific binding studies with cells ............................................................ 81 2.4.7 Applications .............................................................................................................. 81 2.4.7.1 Covalent conjugation and pH sensing............................................................... 81 2.4.7.2 Conjugation with TAC and applications........................................................... 82  Development of single-chain polymer nanoparticles for cellular immunolabeling........................................................................................................................................................84  3.1 Introduction ................................................................................................................... 84  3.2 Results and discussion .................................................................................................. 87 3.2.1 Synthesis of SCPNs .................................................................................................. 87 3.2.2 Photophysical characterization of SCPNs ................................................................ 89 3.2.2.1 Photophysical properties ................................................................................... 89 3.2.2.1.1 Spectra and quantum yield .......................................................................... 89 3.2.2.1.2 Lifetime ....................................................................................................... 91 3.2.2.2 Additional fluorescence measurements ............................................................ 92 3.2.3 Cell viability assay .................................................................................................... 93 3.2.4 Immunolabeling with SCPN-Biotin .......................................................................... 95 3.3 Conclusion .................................................................................................................... 98 3.4 Experimental section ..................................................................................................... 98 xi  3.4.1 Materials ................................................................................................................... 98  3.4.1.1 Preparation of FITC-labeled bovine serum albumin (BSA-FITC) ................... 99 3.4.2 Spectral characterization ........................................................................................... 99 3.4.3 Quantum yield measurements ................................................................................... 99  3.4.4 Fluorescence lifetime measurements ...................................................................... 100 3.4.5 Fluorescence anisotropy measurements .................................................................. 100 3.4.6 Photobleaching measurements ................................................................................ 100 3.4.7 Stern-Volmer quenching measurements ................................................................. 101 3.4.8 Cellular viability assay ............................................................................................ 101 3.4.9 Cellular immunolabeling with SCPN-Biotin (P6) .................................................. 102  Cellular immunolabeling with a polymer dot consisting of a polyacrylate backbone polymer with pendantly functionalized blue HMAT-ODA Dye ..........................103 4.1 Introduction ................................................................................................................. 103  4.2 Results and discussion ................................................................................................ 106 4.2.1 Characterization of HMAT-ODA Pdots ................................................................. 106 4.2.2 Fixed SK-BR3 immunolabeling with HMAT-ODA bioconjugates ....................... 108 4.2.2.1 SK-BR3 immunolabeling with primary anti-HER2 HMAT-ODA Pdot bioconjugate (direct) ....................................................................................................... 108 4.2.2.2 SK-BR3 immunolabeling with secondary goat anti-mouse IgG HMAT-ODA Pdot bioconjugate (indirect) ............................................................................................ 109 4.2.2.2.1 Two-photon confocal fluorescence microscopy ....................................... 111 4.3 Conclusions ................................................................................................................. 112 4.4 Experimental section ................................................................................................... 113 4.4.1 Immunolabeling of fixed SK-BR3 cells with fluorescent nanoparticles ................ 113 4.4.1.1 Preparation of fixed SK-BR3 cells ................................................................. 113 4.4.1.1.1 Cell culture ................................................................................................ 113 4.4.1.1.2 Cell fixation ............................................................................................... 113 4.4.1.2 HMAT-ODA Pdots ......................................................................................... 113 4.4.1.2.1 Cell immunolabeling direct approach ....................................................... 113 4.4.1.2.2 Cell immunolabeling indirect approach .................................................... 114 4.4.1.2.3 Fluorescence microscopy .......................................................................... 114 xii  4.4.1.2.4 Two-photon microscopy ........................................................................... 115  Cellular immunolabeling with semiconducting π-conjugated polymer dots ....116 5.1 Introduction ................................................................................................................. 116  5.2 Results and discussion ................................................................................................ 119 5.2.1 Characterization of F8BT or CNMEHPPV Pdots .................................................. 119 5.2.2 Fixed SK-BR3 immunolabeling with F8BT/CNMEHPPV Pdot bioconjugates via TAC-HER2 ......................................................................................................................... 120  5.3 Conclusion .................................................................................................................. 123 5.4 Experimental section ................................................................................................... 124 5.4.1 Synthesis of Pdots ................................................................................................... 124 5.4.1.1 F8BT and CNMEHPPV .................................................................................. 124 5.4.2 Immunolabeling of fixed SK-BR3 cells with F8BT/CNMEHPPV Dex-Pdots ...... 124 5.4.2.1.1 Preparation of (Anti-HER2-TAC)-Dex-Pdot Conjugates ......................... 124 5.4.2.1.2 SK-BR3 immunolabeling and cell imaging .............................................. 124  Fully self-assembled silica nanoparticle-semiconductor quantum dot supra-nanoparticles and immunoconjugates for enhanced cellular imaging by microscopy and smartphone camera ...................................................................................................................126  6.1 Introduction ................................................................................................................. 126  6.2 Results and discussion ................................................................................................ 129 6.2.1 SK-BR3 immunolabeling with SiO2@(QD-Dex) supra-nanoparticles .................. 129 6.2.1.1 Preparation of SiO2@(QD-Dex) supra-nanoparticles ..................................... 129 6.2.1.2 Characterization of SiO2@QD supra-nanoparticles ....................................... 130 6.2.1.2.1 Physical characterization ........................................................................... 130 6.2.1.2.2 Photophysical characterization .................................................................. 131 6.2.1.3 Immunolabeling with SiO2@(QD-Dex) supra-nanoparticles ......................... 132 6.2.1.3.1 Microscope-based cellular imaging .......................................................... 132 6.2.1.3.2 Smartphone-based cellular imaging .......................................................... 134 6.2.1.3.3 Immunolabeling with multiple colors of SiO2@QD ................................. 135 6.3 Conclusion .................................................................................................................. 136 6.4 Experimental section ................................................................................................... 137 6.4.1 Preparation of SiO2@(QD-Dex) conjugates ........................................................... 137 xiii  6.4.1.1 Preparation of API-modified dextran (API-Dex) ........................................... 137 6.4.1.2 Preparation of supra-NPs ................................................................................ 138 6.4.2 Immunolabeling of fixed SK-BR3 cells with SiO2@(QD-Dex) ............................. 138 6.4.2.1 Preparation of TAC Anti-HER2 complexes ................................................... 138 6.4.2.2 Immunolabeling fixed SK-BR3 cells .............................................................. 139 6.4.3 Imaging SiO2@(QD-Dex) immunolabeled SK-BR3 cells...................................... 139 6.4.3.1 Microscope imaging........................................................................................ 139 6.4.3.2 Smartphone-based imaging ............................................................................. 140 6.4.3.3 Image processing ............................................................................................ 140  Supraparticle assemblies of magnetic nanoparticles and quantum dots for selective cell isolation and counting on a smartphone-based imaging platform ..................141 7.1 Introduction ................................................................................................................. 141  7.2 Results and discussion ................................................................................................ 143 7.2.1 Smartphone-based imaging platform ...................................................................... 143 7.2.2 Composite MNP@QD nanoparticles ...................................................................... 147 7.2.2.1 Necessity of API-modification of MNPs ........................................................ 147 7.2.2.2 Estimating the number of QDs per MNP ........................................................ 148 7.2.2.3 Importance of overcoating with API-Dex....................................................... 149 7.2.2.4 Additional characterization of MNP@QDs .................................................... 150 7.2.2.4.1 SEM-EDX ................................................................................................. 151 7.2.2.4.2 X-ray photoelectron spectroscopy ............................................................. 152 7.2.2.4.3 Infrared spectroscopy ................................................................................ 153 7.2.2.4.4 Nanoparticle tracking analysis (NTA) ...................................................... 154 7.2.2.4.5 MNP@QD and QD optical characterization ............................................. 156 7.2.2.4.6 Fluorescence microscopy imaging of MNP@QDs ................................... 157 7.2.2.4.7 Smartphone imaging of magnetic and PL MNP@QD properties ............. 158 7.2.3 Immunomagnetic cell isolation and imaging .......................................................... 158 7.2.4 Live cell counting ................................................................................................... 161 7.2.5 Discussion ............................................................................................................... 162  7.3 Conclusion .................................................................................................................. 165 7.4 Experimental section ................................................................................................... 166 xiv  7.4.1 Materials ................................................................................................................. 166  7.4.2 Smartphone-based imaging platform ...................................................................... 167 7.4.3 Preparation of API-modified dextran (API-Dex) ................................................... 167 7.4.4 Preparation of MNP@QD. ...................................................................................... 167 7.4.4.1 Epichlorohydrin-crosslinked MNP (X-MNP)................................................. 167 7.4.4.2 Oxidized X-MNP (Ox-MNP) ......................................................................... 168 7.4.4.3 1-(3-aminopropyl)imidazole-modified MNP (API-MNP) ............................. 168 7.4.4.4 Ligand exchange of QDs ................................................................................ 168 7.4.4.5 Self-assembly of MNP@QD .......................................................................... 169 7.4.4.6 Overcoating MNP@QD with API-modified Dextran .................................... 169 7.4.5 MNP@QD characterization .................................................................................... 169 7.4.6 TAC anti-HER2 complexes .................................................................................... 170 7.4.7 Cell-counting assay ................................................................................................. 171 7.4.8 Fluorescence microscopy ........................................................................................ 171 7.4.9 Cell culture .............................................................................................................. 171 7.4.10 PBS buffers ......................................................................................................... 172 7.4.11 Counting DAPI-stained cells on the SIP ............................................................. 172 7.4.12 Counting and imaging fixed SK-BR3 cells labeled with MNP@QDs ............... 173 7.4.13 Counting live SK-BR3 cells on the SIP .............................................................. 174 7.4.13.1 Counting increasing numbers of SK-BR3 cells without  MDA-MB-231 cells ........................................................................................................ 174 7.4.13.2 Counting increasing numbers of SK-BR3 cells with a constant number of MDA-MB-231 cells ........................................................................................................ 175 7.4.13.3 Counting a constant number of SK-BR3 cells with an increasing number of MDA-MB-231 cells ........................................................................................................ 176  Affinity immobilization of semiconducting quantum dots and metal nanoparticles on cellulose paper substrates.....................................................................................................177 8.1 Introduction ................................................................................................................. 177  8.2 Results and discussion ................................................................................................ 179 8.2.1 Immobilization chemistries ..................................................................................... 179 8.2.2 Immobilization of QDs ........................................................................................... 180  xv  8.2.3 Patterning QDs by microcontact printing ............................................................... 182 8.2.4 Immobilization of Au and Pt NPs ........................................................................... 186 8.2.5 Applications with immobilized metal NPs ............................................................. 188 8.2.6 Discussion ............................................................................................................... 189  8.3 Conclusion .................................................................................................................. 190 8.4 Experimental section ................................................................................................... 191 8.4.1 Materials ................................................................................................................. 191  8.4.2 Preparation of paper substrates ............................................................................... 192 8.4.3 Immobilization of QDs and metal NPs ................................................................... 192 8.4.4 Characterization ...................................................................................................... 192  8.4.5 Patterning QDs ........................................................................................................ 193  8.4.5.1 Preparation of molds for PDMS stamps ......................................................... 193 8.4.5.2 Preparation of PDMS stamps .......................................................................... 194 8.4.5.3 Stamping of QDs............................................................................................. 194 8.4.5.4 On-paper ligand exchange .............................................................................. 195  Conclusions and future work ................................................................................197  9.1 Conclusions ................................................................................................................. 197 9.2 Future work ................................................................................................................. 198  9.2.1 Dextran-QDs ........................................................................................................... 199 9.2.2 Antigen profiling with luminescent nanoparticles .................................................. 199 9.2.3 Immunolabeling other cell types with luminescent particles .................................. 202 9.2.4 Upgrading the smartphone-based imaging platform............................................... 202 9.2.5 Direct comparisons with conventional techniques ................................................. 203 9.2.6 Array-based sensing with patterned QDs ............................................................... 204 9.3 Closing remarks .......................................................................................................... 205 References ...................................................................................................................................206  Appendices ..................................................................................................................................236   ............................................................................................................................. 236  A.1 1H NMR .................................................................................................................. 237 A.2 FTIR of ligands ....................................................................................................... 241 A.3 Colorimetric tests .................................................................................................... 241 xvi  A.4 FTIR of Dex-QDs ................................................................................................... 243 A.5 Absorbance, emission, excitation of QDs ............................................................... 243  A.6 Size characterization ............................................................................................... 244 A.7 Interpreting gel electrophoresis results ................................................................... 245  ............................................................................................................................. 248  B.1 Circuit diagram for laser diode controller ............................................................... 248 B.2 Smartphone images of cells isolated with MNP@QDs .......................................... 248 B.3 Image analysis for cell counting ............................................................................. 249 B.4 Signal-to-noise ratio calculation ............................................................................. 251 B.5 Comparison to other smartphone-based cell counting systems .............................. 252  xvii  List of Tables  Table 1.1 Summary table for applications of QD immunolabeling. ............................................ 29 Table 1.2 Representative examples of smartphone-based point-of-care diagnostics................... 43 Table 2.1 Solvodynamic sizes of X-QD600. Data contributed by Kelly Rees. ........................... 57 Table 3.1 Photophysical properties of SCPN–P7 and reference materials. ................................. 89 Table 4.1 Fluorescence properties of HMAT-ODA Pdots. ........................................................ 108 Table 5.1 Fluorescence properties of F8BT and CNMEHPPV Pdots........................................ 120 Table 5.2 Fluorescence microscopy optics for SK-BR3 labeling. ............................................. 125 Table 7.1 Fluorescence microscopy optics for SK-BR3 labeling. ............................................. 174 Table 8.1 Specification for stamp molds (in AutoCAD) ........................................................... 194  xviii  List of Figures  Figure 1.1. Jablonski diagrams illustrating the processes of photon absorption, fluorescence, and competing non-radiative processes. The y-axis is potential energy and the x-axis represents the nuclear coordinate. Singlet (S୬) and triplet (T୬) electronic states are shown as potential energy wells, each with vibrational states (ν୬) superimposed. (A) The processes involved in a fluorescence relaxation pathway: (i) photon absorption results in (for example) electron promotion from S଴ν଴ to Sଵνଶ; (ii) the electron transitions to Sଵν଴ by vibrational relaxation; and (iii) radiative relaxation to the ground electronic state, generating a photon. (B) Processes involved in a non-radiative relaxation pathway. Photon absorption results in (for example) a transition from S଴ν଴ to (i) Sଵνଶ or (ii) Sଶνଶ. From Sଶνଶ, vibrational relaxation transitions the electron to Sଶν଴, where (iv) internal conversion to  Sଵν଻ occurs. Next, vibrational relaxation to Sଵν଴ occurs, after which (v) internal conversion to S଴νଵଷ is followed by vibrational relaxation to S଴ν଴. (C) Other relaxation pathways: (i) photon absorption excites an electron from S଴ν଴ to Sଵνଶ, followed by (ii) vibrational relaxation to Sଵν଴. Next, (iii) intersystem crossing to Tଵνଵ occurs and is followed by (ii) vibrational relaxation to Tଵν଴. From Tଵν଴, radiative relaxation can occur via phosphorescence (not shown), reverse intersystem crossing and delayed fluorescence (not shown), or by non-radiative pathways similar to B. ................................................................................................................................................. 3  Figure 1.2. An illustration of the Franck-Condon principle for electronic transitions from the ground electronic state (S଴) to an excited state (Sଵ). (A) The absorption/emission intensity of an electronic transition is proportional to the overlap between the initial and final vibronic wavefunctions. The blue line represents the most probable transition from S଴ν଴ to Sଵνଶ, which has the greatest overlap between vibronic wavefunctions. The orange line represents the most probable transition from Sଵν଴ to S଴νଶ. The probability density functions are shown in the shaded areas of the wavefunctions. (B) Corresponding absorption and emission spectra and possible transitions from panel A. Note that the individual transitions are typically only observed in the gas phase or at very low temperatures. In solution, around room temperature, the spectral features are usually broadened into a smooth shape (shaded regions). As per the FCFs, the absorption and emission spectra are mirror images of each other. ......................................................................................... 6  xix  Figure 1.3 Primary (direct) versus secondary (indirect) immunolabeling strategies. A nanoparticle (NP) label is conjugated with either a primary or secondary antibody. The primary antibody binds specifically to the cell antigen, whereas the secondary antibody binds to a primary antibody. ... 15 Figure 1.4. Overview of QDs. (A) (i) Atomistic illustration of two sizes of QD nanocrystal. (ii) High-resolution TEM image of a QD. Reprinted with permission from ref.28 (B) Absorption and fluorescence spectra for various sizes of QDs. Reprinted with permission from ref.37 (C) Size-tunable PL of CdSe QDs. The photograph was taken under UV illumination (365 nm). Reprinted with permission from ref. 28 (D) Approximate wavelength ranges over which the fluorescence of various QD materials can be tuned through control of nanocrystal size. The visible spectrum is between ca. 400–650 nm. Reprinted with permission from ref. 29 ............................................... 20 Figure 1.5 Common bioconjugation strategies with peptides, proteins, and antibodies. (a) Coupling between thiol and amine groups using SMCC. (b) Coupling between carboxyl and amine groups using EDC activation (and often NHS, not shown). (c) Hydrazone coupling between hydrazide and aldehyde groups. (d) Binding of polyhistidine-tagged peptides (and proteins, not shown) to the inorganic shell of ligand-coated QDs. QDs can also be functionalized with coatings that display Ni2+-nitrilotriacetic acid groups for binding polyhistidine-tagged peptides (not shown). (e) Binding between StreptAvidin-QD conjugates and biotinylated antibodies. Figure adapted with permission from ref. 62. Copyright 2007 Nature Publishing Group. ............................................. 25 Figure 1.6 Common bioconjugation strategies with aptamers and other oligonucleotides. (A) Coupling between carboxyl and amine groups using EDC activation (and often NHS, not shown). (B) Coupling between thiol and amine groups using SMCC. (C) Binding of a dithiol-terminated aptamer to the inorganic shell of ligand-coated QDs. (D) Binding of a biotinylated aptamer to StreptAvidin-QD conjugates. (E) Binding of polyhistidine-tagged aptamer to the inorganic shell of ligand-coated QDs. Parts of the figure have been adapted with permission from ref. 62. Copyright 2007 Nature Publishing Group. .................................................................................................... 26  Figure 1.7 (A) Simplified schematic of a CMOS image sensor. (B) Spectral sensitivity of a typical CMOS image sensor without (black) and with RGB colour filters (coloured lines). The typical blocking region of an IR filter is also shown. Figure adapted with permission from ref.111 ........ 41 Figure 1.8 Luminescent nanoparticles evaluated for immunolabeling: Single Chain Polymer Nanoparticles (SCPN), HMAT-ODA Pdot, F8BT Pdot, CNMEHPPV Pdot, SiO2@QD, and MNP@QD. Approximate hydrodynamic diameters: SCPNs 26 nm, Pdots 50–70 nm, QDs 2–10 xx  nm, SiO2@QDs ~110 nm, MNP@QDs ~250 nm. The inset table shows the approximate brightness values for the luminescent nanoparticles. Brightness values are approximations based on the calculated product of quantum yield and molar absorption coefficient for the respective materials, with excitation at 405 nm for all materials except SCPNs (450 nm). The ranges for the brightness of QDs and Polymer dots are a function of nanoparticle size and material. ................................. 44  Figure 1.9 Strategies for immunolabeling fixed SK-BR3 cells with fluorescent nanoparticles. . 45 Figure 2.1 (A) General structures of the dextran ligands. The R group may be another modification, an unreacted aldehyde (or the corresponding hydrate), or have cyclized with the secondary amine of the shown modification. (B) Cartoon schematics (left) and scale illustrations (right) of the two general QD-functionalization strategies: terminal (reducing end) modification of dextran with an anchoring group, and pendant modification of dextran with multiple anchoring groups. For the illustrations, the translucent sphere around the QD approximately corresponds to the measured hydrodynamic radius, rH. Multiple possible arrangements for the pendant-functionalized dextran are shown. (C) Cartoon schematic of a TAC (left) and scale illustration (right) for a TAC bound to a dextran-functionalized QD. Scale illustrations contributed by Katherine Krause. ......................................................................................................................... 52  Figure 2.2 Anthrone assay for determining the number of dextran ligands per QD. The values are calculated based on quadruplicate measurements and are expressed as the mean (± 1 standard deviation). ..................................................................................................................................... 54  Figure 2.3 Confirmation of dextran-functionalization of QDs by selective aggregation with ConA. Top: Photographs of X-QD solutions without (–) ConA, with (+) ConA, and with both ConA and glucose (Glc). The white arrows indicate aggregates. DHLA-QDs were used as a non-dextran control and Glc was used for competitive binding with ConA. Bottom: Agarose gel electrophoresis of analogous samples. The Dex-QD samples with ConA (and without Glc) aggregated and did not migrate from the wells. The dashed white line represents the well positions. The solid white line represents a gap of dark space digitally edited from the gel. The brightness of the DHLA samples (orange dashed box) was digitally enhanced for better visibility in the figure. Data was contributed by Kelly Rees. ............................................................................................................................... 55  Figure 2.4 (D10-t-DHLA)-QD600 incubated with 100 molar equivalents of dextran-specific (ConA) and non-specific (BSA, Lyz) proteins. Wells are labelled with a white dashed line. Agarose xxi  gels were prepared as a 0.5% (w/v) solution in TBE buffer. Gels were run at ~6.7 V cm–1 for 30 min. ............................................................................................................................................... 56  Figure 2.5 (A) Agarose gel of His-QDs and QD samples prepared using precursors to the fully-modified dextran ligands (i.e. no DHLA or API groups). (B) Agarose gel of His-QDs, Dex-QDs prepared with the fully-modified dextran ligands, and DHLA-QDs as a reference. (C) Electropherograms of the different QD samples with fluorescein as an anionic internal standard (eluted at ~4 min). Rhodamine B was used a neutral reference dye marker (eluted at ~2 min). The fluorescein internal standard is indicated with a black arrow. The elution times for the QDs and Rhodamine B are indicated with the coloured, horizontal arrows. Data collected in collaboration with Kelly Rees ............................................................................................................................. 60  Figure 2.6 (A) Photos of samples of (D10-t-DHLA)-QD600 and DHLA-QD600 in buffers at various pH after initial preparation and after 8 weeks. The photos show the QD PL under UV illumination. The white arrows indicate the presence of visible aggregation. (B) Summary plot of the approximate period of colloidal stability at pH 3–8 and at 1 M ionic strength. Data contributed by Kelly Rees. ............................................................................................................................... 61  Figure 2.7 Agarose gels of X-QDs incubated with 1.0, 5.0, and 9.5 mg/mL protein solutions (in carbonate buffer, excluding plasma, which is percent v/v), where X = D10-t-DHLA, D6-p-DHLA, D6-p-API, and DHLA. Strong non-specific protein binding was observed with DHLA-QDs. Depending on the protein(s), some Dex-QDs exhibited little or no non-specific binding, whereas others exhibited some non-specific binding. Refer to text for details. Data contributed by Melissa Massey. ......................................................................................................................................... 63  Figure 2.8 (A) Representative images of live A549 cells that had been incubated with X-QD600 (50 nM), where X = GSH or D10-t-DHLA. Negative controls were cells that were not incubated with QDs. Scale bars are 50 µm. A pixel intensity calibration bar is provided. All images were acquired under the same microscope, camera, and image processing settings. Additional images are shown in Figure 2.9. (B) Average post-wash PL intensity of live A549 and MDA-MB-231 cells incubated with various X-QD samples at 50 nM for 30 min. The average cell autofluorescence (autoFL) was 232 ± 0.7 a.u. and is indicated by the black dashed line and diagonally crosshatched bars. (C) Quantitative assay of non-specific binding of X-QD to live A549 and MDA-MB-231 cells incubated with varying concentrations of X-QD600, where X = GSH, D6-p-DHLA, and D10-t-DHLA. Experiments and data analysis performed with Kelly Rees. ......................................... 65 xxii  Figure 2.9 Comparison of non-specific binding between live cells and QD600 (50 nM) with various ligand coatings. Images were acquired under brightfield and fluorescence modes. (A) A549 cells. (B) MDA-MB-231 cells. Negative controls are cells that were not incubated with QDs. Scale bar = 50 µm. Exposure time = 150 ms. Images were acquired at the same microscope and camera settings. A pixel intensity calibration bar is provided on the right. Experiments and data analysis performed with Kelly Rees. ............................................................................................ 66 Figure 2.10 (A) Brightfield, PL, and merged images of X-QD600 samples injected into the cytosol of live A549 cells, where X = D10-t-DHLA, D6-p-DHLA, and D6-p-API. These images were acquired 30 min post-injection. The QDs dispersed throughout the cytosol but were excluded from the cell nuclei and other intracellular compartments. (B) Brightfield, PL, and overlay images of DHLA-QD600 samples after attempted microinjections (arrows) into A549 cells. These images were acquired 3 min post-injection. All scale bars = 25 µm. The pixel intensity calibration bars for the PL images are provided at the bottom. Data contributed by Hyungki Kim. ........................... 67  Figure 2.11 A549 cell viability assay with (D10-t-DHLA)-QD600 at concentrations between 10 pM and 1 μM. Cell viability is expressed as a percentage of the negative control (cells that were not incubated with QDs). Data points and error bars are the average and standard deviation of three replicates. The trendline is only to guide the eye.......................................................................... 68  Figure 2.12 (A) Schematic for the reaction between dextran (as ligands on a QD, not shown) and FITC. (B) PL image of an agarose gel of FITC-labeled Dex-QDs and the PL emission spectra measured for each of the gel bands (labeled with Roman numerals). .......................................... 69  Figure 2.13 Emission spectra for (D6-t-DHLA)-QD600 (left) and FITC (right) at various pH values. ........................................................................................................................................... 69  Figure 2.14 Proof-of-concept pH sensing with FITC-(D6-t-DHLA)-QD600 (clockwise): (i) photograph of a well-plate loaded the FITC-labelled QDs in buffers at different pH; (ii) PL emission spectra of the FITC-(D6-t-DHLA)-QD600 at in buffers at various pH; (iii) plot of the FITC/QD600 PL emission ratio as a function of pH. ................................................................... 70 Figure 2.15 EPO immunofluorescent assay with Dex-QD TAC conjugates. (A) Cartoon schematics and agarose gels of Dex-QDs binding with (i) anti-dextran antibody and (ii) a full TAC. Binding is indicated by electrophoretic mobility shifts and increased streaking. (B) Proof-of-concept sandwich immunoassay for EPO: (i) cartoon schematic of the assay format with (D10-t-DHLA)-QD645 and (ii) measured PL intensity for a sample with EPO (5 mU) and without EPO. xxiii  Error bars are the average of three replicates. The contrast ratio was 8:1, with a p-value of 0.04. Data contributed by Melissa Massey. ........................................................................................... 71  Figure 2.16 Cellular immunofluorescent labelling with TAC conjugates of Dex-QDs. (A) Immunolabeling of HER2-positive SK-BR3 breast cancer cells: (i) cartoon schematic of the experiment; (ii) representative images of cells incubated with (D6-p-DHLAm)-QD600 with and without TAC-Anti-HER2 (scale bar = 80 µm in main images, 20 µm in inset); and (iii) average PL intensity measured for cells with (D6-p-DHLAm)-QD600 with and without TAC-Anti-HER2 (error bars are the standard deviation from imaging 11 cells). Images were acquired with the same microscope, camera, and image processing settings. .................................................................... 71 Figure 3.1 Cartoon schematic of SCPNs. Approximate hydrodynamic diameter: 26 nm. .......... 86 Figure 3.2 Strategy for immunolabeling fixed SK-BR3 cells with SCPN-Biotin. ...................... 87 Figure 3.3 SCPN-forming polyacrylamides. The polymer chain collapsing agent (BTA-NH2) is in orange. The fluorescent dye (FITC) is in green. P5 is the precursor polymer to P6. The cell-labeling bioconjugate handle (biotin) for P6 and cell-targeting molecule (FA) for P7 are in pink. The cytotoxic payload (CPT) for P4 is in red. Daniel Bajj synthesized the polymers. ....................... 88 Figure 3.4 Photophysical characterization of SCPN-P7 and reference materials. (A) Spectra: (i) absorption; (ii) fluorescence excitation; and (iii) fluorescence emission. (B) Fluorescence decays. (C) Photobleaching curves. (D) Stern-Volmer plots for collisional quenching by iodide ion. All measurements were made in 1 PBS buffer at pH 7.2 (see section 3.4.1 for recipe) between 21–26 C. Error bars represent the standard deviation of three replicate measurements. The concentrations for each sample can be found in the section 3.4. Data collected and analyzed in collaboration with Hsin-Yun Tsai and Hyungki Kim. .................................................................. 93  Figure 3.5 SK-BR3 cell viability assay with CPT-SCPN-N3 (P4), and CPT-SCPN-FA (P7). The nanoparticle concentrations were from 10 pM–16 µM (810–7–1.3 mg/mL) in PBS buffer (see Section 3.4 for recipe) and incubated with cells at 37 C. Cell viability is expressed as a percentage of the negative control (cells that were not incubated with the SCPN). Data points and error bars are the average and standard deviation of three replicates. The dashed lines are to guide the eye. The solid trendlines show that viability does not decrease as the concentration of SCPN increases........................................................................................................................................................ 95  Figure 3.6 Fluorescent labeling of SK-BR3 breast cancer cells with SCPN-Biotin (P6). Top row: brightfield images. Middle row: fluorescence images (the pixel intensity calibration scale is xxiv  indicated on the right). Bottom row: merged brightfield and fluorescence images. The sample is in the right-most column. The “+” indicates that the cells were incubated with the respective material. Scale bars = 100 µm...................................................................................................................... 96  Figure 3.7 Fluorescent labeling of SK-BR3 breast cancer cells with SCPN-Biotin (P6). Top row: brightfield images. Middle row: fluorescence images. Bottom row: merged brightfield and fluorescence images. The sample is in the left-most column and negative controls are in the other three columns. The “+” indicates that the cells were incubated with the respective material. Scale bars = 100 µm. This figure represents a lower contrast example of the experiment shown in Figure 3.6, performed several weeks later using the same stock solution. .............................................. 97 Figure 4.1 Preparation of antibody conjugated HMAT-ODA Pdots for cellular immunofluorescent labeling. Approximate hydrodynamic diameters: Pdots 50–70 nm. (A) Structure of blue-emitting HMAT-ODA polymer, which is an acrylate copolymer of phenylcarbazole (grey circle) and HMAT-ODA (green, and orange circles). Chris Morty Tonge synthesized the polymer. (B) Structure of amphiphilic polymer poly(styrene)-graft-poly(ethylene oxide) functionalized with carboxylic acid (PS-PEG-COOH). (C) Illustration of HMAT-ODA Pdot, Kelsi Lix synthesized the Pdots. (D) Illustration of goat anti-mouse IgG conjugated HMAT-ODA Pdot. Rupsa Gupta prepared the conjugates............................................................................................................... 105  Figure 4.2 Strategies for immunolabeling fixed SK-BR3 cells with HMAT-ODA Pdots. ....... 106 Figure 4.3 Nanoparticle tracking analysis of native HMAT-ODA/PS-PEG-COOH Pdots (black) and goat-anti-mouse IgG conjugated HMAT-ODA/PS-PEG-COOH Pdots (blue). Data contributed by Rupsa Gupta. .......................................................................................................................... 107  Figure 4.4 Immunolabeling of fixed, SK-BR3 cells with HMAT-ODA@anti-HER2 and HMAT-ODA (control) Pdots. Scale bars = 50 μm. A normalized pixel intensity calibration bar is provided for the relative HMAT-ODA fluorescence intensity. Images were acquired under the same microscope and camera settings. The images were adjusted to the same brightness contrast. The monochrome images were pseudo-colored cyan. ....................................................................... 109 Figure 4.5 Differential-interference contrast (DIC, top row) and fluorescence (bottom row) microscopy images of SK-BR3 cells labeled with HMAT-ODA-Pdot-Goat anti-mouse IgG with (+) and without (─) primary labeling using anti-HER2. The images were acquired under the same microscope and camera settings. The images were adjusted to the same brightness and contrast. The monochrome images were pseudo-colored cyan. ................................................................ 110  xxv  Figure 4.6 Two-photon excitation fluorescence images of SK-BR3 cells labeled with primary anti-HER2 followed by HMAT-ODA-Pdot-Goat anti-mouse IgG (left) and without primary anti-HER2 (right). The dashed white circle represents the approximate shape and location of the cell. The images represent one focal plane within in the cells. Magnification: 63×. Excitation: 770 nm (2PE). Emission: 660 nm short-pass dichroic mirror. Scale bar = 25 μm. ............................................. 112 Figure 5.1 Preparation of dextran-functionalized F8BT or CNMEHPPV Pdots for immunofluorescent cellular labeling. Approximate hydrodynamic diameters: Pdots 50–70 nm. Pdots were synthesized by Kelsi Lix. ......................................................................................... 118  Figure 5.2 Strategies for immunolabeling fixed SK-BR3 cells with dextran-coated F8BT and CNMEHPPV Pdots. .................................................................................................................... 119  Figure 5.3 Immunolabeling of fixed, and DAPI (nuclear stain) labeled SK-BR3 cells with F8BT/Dex Pdots via TAC-HER2 (bottom row). Scale bars = 50 μm. A normalized pixel intensity calibration bar (below) is provided for the relative F8BT fluorescence intensity. Images were acquired under the same microscope and camera settings. The images were adjusted to the same brightness contrast for the different color channels. The monochrome images were pseudo-colored green (F8BT) or blue (DAPI). .................................................................................................... 121  Figure 5.4 Comparison of fixed SK-BR3 immunolabeling with dextran-functionalized (F8BT/Dex) and un-modified (F8BT) Pdots via TAC-HER2. A normalized pixel intensity calibration bar (below) is provided for the relative F8BT fluorescence intensity. Images were acquired under the same microscope and camera settings. The images were adjusted to the same brightness contrast for the different color channels. The monochrome images were pseudo-colored green (F8BT). .............................................................................................................................. 122  Figure 5.5 Immunolabeling of fixed, and DAPI (nuclear stain) labeled SK-BR3 cells with CNMEHPPV//Dex Pdots via TAC-HER2 (bottom row). Scale bars = 50 μm. A normalized pixel intensity calibration bar (below) is provided for the relative CNMEHPPV fluorescence intensity. Images were acquired under the same microscope and camera settings. The images were adjusted to the same brightness contrast for the different color channels. The monochrome images were pseudo-colored red (CNMEHPPV) or blue (DAPI). .................................................................. 123  Figure 6.1 Scheme for the preparation of self-assembled SiO2@QD supra-nanoparticles and their functionalization with dextran. SiO2 NPs are modified with triethoxy-3-(2-imidazolin-1-yl)propylsilane (IPS), where the imidazoline groups spontaneously bind (electrostatically or xxvi  coordinate) CdSe/CdS/ZnS QDs coated with glutathione ligands (not shown). The QDs are then further functionalized with dextran that has been modified with pendant 1-(3-aminopropyl)imidazole groups. The SiO2 NPs and QDs are drawn approximately to scale. The IPS and dextran molecules are not drawn to scale. The SiO2@QD were prepared by Ghinwa Darwish and Jérémie Asselin. ................................................................................................................... 128  Figure 6.2 Strategies for immunolabeling fixed SK-BR3 cells with dextran-coated SiO2@QDs...................................................................................................................................................... 129  Figure 6.3 TEM image of SiO2@QD supra-nanoparticles. The inset is a higher magnification image. Scale bar represents 100 nm. Inset scale bar represents 50 nm. TEM images contributed by Ghinwa Darwish and Jérémie Asselin. ....................................................................................... 131  Figure 6.4 (A) TAC-based immunolabeling of HER2-expressing SK-BR3 cells with Dex-QD605, SiO2@(QD605-GSH), and SiO2@(QD605-Dex), and control experiments without TAC. Brightfield images are shown alongside DAPI (nuclear stain) fluorescence images and QD PL images. A pixel-intensity calibration bar is shown. Scale bar = 20 µm in all images. (B) Comparison of the contrast ratios for specific labeling (with TAC) and non-specific binding (without TAC) for SiO2@(QD605-GSH) and SiO2@(QD605-Dex). For comparison, 10 cells were analyzed for each sample. (C) Images highlighting the difference in signal-to-background ratio between specific labeling (with TAC) with SiO2@(QD605-GSH) and SiO2@(QD605-Dex). The scale bars = 100 µm and a pixel-intensity calibration bar is shown. .......................................... 133  Figure 6.5 (A) Cross-sectional view for the design of smartphone-based platform for PL imaging of cells. Details of the design can be found in Chapter 7. (B) Smartphone-acquired images of SK-BR3 cells that were TAC-immunolabeled with Dex-QD605, SiO2@(QD605-GSH), and SiO2@(QD605-Dex), and control experiments without TAC. Scale bars = 500 μm. ................ 134 Figure 6.6 Smartphone-based imaging of fixed SK-BR3 cells immunolabeled with SiO2@(QD-Dex) supra-nanoparticles without (─) TAC-HER2, and with (+) TAC-HER2.The QD colors are denoted as QD λ, where λ represents the center peak wavelength of the QD. Scale bars = 500 μm...................................................................................................................................................... 136  Figure 7.1 (A) Schematic of an MNP@QD supraparticle assembly: QDs are bound to an imidazole-modified dextran coating on the MNPs and overcoated with additional imidazole-modified dextran. (B) Schematic of TAC-mediated binding of an MNP@QD to HER2 antigen on the surface of an SK-BR3 cell and isolation by magnetic pull-down. (C) Zoomed view of the TAC-xxvii  mediated binding. (D) Diagram illustrating the steps in the cell counting assay: TAC and MNP@QD are added to a sample cell suspension and the target cells are pelleted magnetically, washed, resuspended, and transferred to a chamber slide for enumeration on the SIP. The assay is demonstrated with a mixture of HER2-positive and HER2-negative breast cancer cells. ......... 143 Figure 7.2 (A) Rendering of the smartphone-based fluorescent cell imaging platform (SIP). (B) Optical design for imaging (L1, ƒ = 25 mm plano-convex; L2, ƒ = 19 mm achromatic doublet lens; emission filter). The laser is 20 mW at a wavelength of ~405 nm. The laser beam-shaping optics are omitted for clarity but can be found in Figure 7.3. (C) Photo of the SIP with the smartphone mounted. The thumb-screw stage-adjusters are used for manual focusing. The laser-adjustment dial modifies the angle and lateral position of the laser. The smartphone powers the laser diode, the output intensity of which can be adjusted with the rheostat. (D) Photo of the SIP with top stage removed for a top-down view, with the sample slide illuminated by the laser. ......................... 144 Figure 7.3. Laser diode optics. The dispersed light from the laser diode is focused through an aspheric lens (ƒ = 3.3 mm) and a plano-convex lens (ƒ = 15 mm), then diverged by a plano-concave lens (ƒ = –15 mm). ...................................................................................................................... 145  Figure 7.4 (A) Comparison of fixed and DAPI-stained SK-BR3 cell-counting on the SIP and a commercial Countess cell counter. The Countess II cell counter enumerates the cells by imaging the samples in brightfield mode in solution and applying their cell-counting image analysis algorithm. (B) SIP images of DAPI-stained SK-BR3 cells at various concentrations. .............. 146 Figure 7.5 (A) Comparison of GSH-QD605 binding to API-MNP and MNP. PL emission microscopy of API-MNP and MNP after mixing with GSH-QDs. A pixel intensity bar is shown for reference. Scale bar = 200 µm. The out-of-scale white coloring is for illustrative purposes; the PL signal did not saturate the detector. Quantitative analysis was not affected by this scaling. (B) The measured PL intensity for each sample was measured on a spectrometer that was coupled to the trinocular head of the microscope. The MNP concentrations, as measured by NTA, were kept the same............................................................................................................................... 147  Figure 7.6. UV-visible extinction spectra, deconvoluted contributions, and fits for estimating the number of QDs assembled per MNP for (A) QD635 and (B) QD575. ...................................... 149 Figure 7.7. Colloidal stability of MNP@QD overcoated with API-modified dextran (API-Dex). (A) Pellet of MNP@QDs shortly after initial preparation. (B) Flocculation of MNP@QDs when incubated with unmodified dextran. (C) Colloidally stable MNP@QD overcoated with API-Dex xxviii  after the same incubation period. (D) As-prepared MNP@QD after pelleting the particles, followed by removing the supernatant, and then washing the pellet with fresh carbonate buffer. The particles were resuspended in fresh buffer and then allowed to incubate on the benchtop for 5 min. MNP@QD incubated with either (E) unmodified dextran or (F) API-Dex. All photographs were taken under ambient lighting conditions. The arrows indicate sample pelleting or flocculation. ................................................................................................................................ 150  Figure 7.8 (A) TEM images of API-MNP and (B) MNP@QD605 (9.8 ± 1.3 nm diameter for QD605). Insets show zoomed views. Inset scale bars are 50 nm. Full image scale bars are 200 nm...................................................................................................................................................... 151  Figure 7.9 (Left) EDX analysis of API-MNP, MNP@QD, and sample substrate (control). Inset graphs represent the 2ꟷ4 keV range. Red-dotted lines align with S and Cd peaks. (Middle, Right) SEM-EDX point analysis of API-MNP and MNP@QD. Spots represent different regions of the SEM images that were analyzed by EDX. In the MNP@QD EDX figure, Spot 4 and 5 were chosen to represent the MNP@QD, and Spots 6 and 7 provided analogous spectra (not shown). Inset graphs represent the 3–4 keV range. Likewise, Spot 1 was chosen to represent the background and substrate, and Spots 2, 3, and 8 provided analogous spectra (not shown). ................................. 152 Figure 7.10. XPS elemental analysis of MNP, API-MNP, and MNP@QD (no API-Dex overcoating). Peaks are labelled according to the element. Peaks corresponding to QDs are labelled in red (Zn, Cd, S). Peaks corresponding to MNPs are labelled in black (Fe, O, C). Peaks corresponding to the glass background (Si) and potassium or sodium salts are labelled in grey...................................................................................................................................................... 153  Figure 7.11. Stacked ATR-IR spectra for API-MNP, MNP@QD, and GSH-QD. Fe-O, C-O, C-H, and O-H bond stretching vibrations are labelled where applicable. ........................................... 154 Figure 7.12 (Left) API-MNP and MNP@QD575 size distributions (252 ± 117 nm, 250 ± 81 nm) determined by NTA. The API-MNP had no measurable PL emission. (Middle, right) NTA sizing data for MNP@QD605 and MNP@QD635. .............................................................................. 155 Figure 7.13. Plots of intensity versus size (top row; scatter plots) and concentration versus intensity (bottom row; histograms) for MNP@QD575, MNP@QD605, and MNP@QD635. .. 156 Figure 7.14 (A) UV-visible extinction (Ext) and PL emission (Em) spectra for API-MNP and MNP@QD575. The arrow highlights the first exciton peak of the QD575. (B) PL excitation spectra xxix  for various colors of MNP@QD. (C) PL excitation spectra for the corresponding His-QD.The QD colors measured were  = 575 nm, 605 nm, and 635 nm. ................................................... 157 Figure 7.15 Fluorescence microscopy image of MNP@QD575 (left; scale bar = 5 µm). A three-dimensional pixel-intensity plot for the boxed area of the image is provided (right). Single MNP@QD assemblies are indicated by the white arrows. ......................................................... 157 Figure 7.16 Magnetic pelleting of MNP@QDs from a colloidal suspension. ........................... 158  Figure 7.17 Smartphone (main images, scale bar = 200 µm) and microscope images (insets, scale bar = 20 µm) of fixed SK-BR3 cells isolated with MNP@QDs of various colors: QD485, QD575, QD605, and QD635, where the notation QDλ refers to the wavelength of peak PL emission for the QD. The smartphone images were acquired in RGB color format. Microscope images are pseudo-colored from the measured monochrome intensity values. ........................................................ 159 Figure 7.18 (A) Differential interference contrast (DIC) and PL images of labeled SK-BR3 cells acquired with a research-grade microscope for different materials combinations. Controls were unlabeled SK-BR3 cells (the cell is outlined with a dashed circle). Scale bars = 20 µm. All images were acquired under the same microscope and camera settings. PL images are false-colored from the measured monochrome intensity values. (B) Average PL signal from labeled cells and surrounding background for each materials combination. A minimum of 20 cells were analyzed from images acquired at a lower magnification (10×). ............................................................... 160  Figure 7.19 (A) SIP images of increasing numbers of SK-BR3 cells isolated with MNP@QDs. A single cell is indicated with the arrow for the 100-cell sample. Scale bar = 200 µm. Each image represents an area of 2 × 3 mm2. (B) Quantification of an increasing number of live SK-BR3 cells (HER2+). (C) Quantification of an increasing number of live SK-BR3 cells in samples spiked with a background of 6140 ± 470 live MDA-MB-231 cells (HER2ꟷ). (D) Quantification of a constant level of live SK-BR3 cells with increasing background of live MDA-MB-231 (HER2ꟷ) cells. All samples were in separation buffer. Additional smartphone images for the assays in panels B-D are shown in Figure B-2, Figure B-3, Figure B-4. ......................................................................... 162 Figure 8.1 Summary of the chemical modifications of cellulose (1): oxidation to yield aldehydes (2) enabled modification with imidazole groups (3) or dithiolane groups (4a). The latter were reduced to dithiol groups (4b). Cellulose (1) was also modified with APTES (5) and, subsequently, dithiolane (or dithiol) groups (6a and 6b). The R group may be another modification, an unreacted aldehyde (or the corresponding hydrate), or may have cyclized with the secondary amine of the xxx  shown modification. The chemistries were designed by Eleonora Petryayeva, who also contributed the XPS, IR, and colourimetric characterization data. ................................................................ 179 Figure 8.2 QDs immobilized on cellulose substrates (3a), (4b), and (6b). (A) 2PE PL images and (B) SEM images (secondary electron mode) at three different magnifications. The insets in the SEM images show the paper substrate without immobilized QDs and are at the same scale as the main images (and thus show a smaller region of interest). (C) PL spectra of modified cellulose substrates after exposure to aqueous QDs and washing. Errors bars are the standard deviation of at least three replicates and are shown only at the peak PL wavelength. Immobilization of multiple colors of QD. (D) Color photographs (smartphone camera) of selected mixtures of red-, green-, and blue-emitting QDs immobilized on circular (3 mm diameter) paper substrates (3). A blank substrate is also shown and has some background blue intensity. The numbers are sample identifiers. (E) Plot of the color of net PL emission for the samples from panel D (open circles) on the CIE 1931 color diagram. Data contributed by Eleonora Petryayeva and Hyungki Kim. ..... 181  Figure 8.3. Patterned immobilization of QDs by microcontact printing. (A) Dot pattern of TOPO-QD605 (left) and crossed-lined patterns of TOPO-QD525 and TOPO-QD605 (right). Images were acquired without washing. (B) Comparison of patterning TOPO-QD605 on substrates (1) and (3), with and without washing post-stamping. Modification of the cellulose substrate was necessary for retention of the pattern. ......................................................................................................... 182  Figure 8.4 Preparing line patterns. Optical image and example of a line profile plot for the 3-D printed mold (left), PDMS stamp (middle), and resulting alkyl-QD605 line pattern on (3) (right)...................................................................................................................................................... 183  Figure 8.5 Image (left) and example of a line profile plot (right) for a dot array of alkyl-QD605...................................................................................................................................................... 183  Figure 8.6 GSH concentrations (determined by Ellman’s assay) for substrate (3), with and without immobilized TOPO-QD605, after exposure to a solution of GSH. The data shown was background subtracted from the signals from the respective control samples (i.e. (3) + TOPO-QD, or (3) alone)...................................................................................................................................................... 184  Figure 8.7 PL micrographs for line patterns (pre-wash) made by stamping GSH-QD605 (top row) and His-QD605 (bottom row) on paper substrates (3). The scale bars are 500 μm.................... 185 Figure 8.8 PL micrographs for line patterns (post-wash) made by stamping His-QD605 on paper substrates (1), (5), and (6a). The scale bars are 500 μm. ............................................................ 186 xxxi  Figure 8.9 PL micrographs for line patterns (post-wash) made by stamping GSH-QD605 on paper substrates (1), (5), and (6a), with and without glycerol in the stamping solution. The scale bars are 500 μm. ....................................................................................................................................... 186  Figure 8.10 Au NPs immobilized on substrates (3a), (4b), and (6b). (A) Photo-micrographs of the substrates without Au NPs (bottom/left) and with Au NPs (top/right). Images contributed by Hyungki Kim. (B) SEM images at three different length scales. The insets show the paper substrate without immobilized Au NPs and are at the same scale as the main images (and thus show a smaller region of interest). Note that the highest-magnification image for (4a) shows two regions of interest with Au NPs. SEM images for all substrates without Au NPs and for (6a) + Au NPs were acquired in secondary electron mode. SEM images for (3) and (4a) with Au NPs were acquired in back-scattered electron mode. (C) Photographs of paper substrates (diameter ~3 mm) with immobilized Au NPs (insets) and corresponding extinction spectra. Note that the extinction spectra are not corrected for the scattering contribution from the paper substrate. Spectra and images contributed by Hyungki Kim. ........................................................................................................................ 187  Figure 8.11 (A) Putative mechanism for the Pt NP-catalyzed decolorization of methyl orange (MO) by sodium borohydride. (B) Methyl orange decolorization kinetics for (4b) with and without Pt NPs. The absorbance was measured at 466 nm and normalized to an initial value of unity. (C) Illustration showing the experimental format for SERS measurements. (D) Superposition of Raman spectra obtained from different Au NP-modified substrates. Data contributed by Hyungki Kim. ............................................................................................................................................ 189  Figure 8.12 AutoCAD renditions (left), and photographs and micrographs of 3D-printed molds (middle) and final PDMS stamps (right). Line array (top) and dot array (bottom) patterns are shown. Photograph scale bars represent 2 mm. Inset micrograph scale bars are 500 μm. ......... 193 Figure 8.13 General microcontact printing method for forming QD-patterns on paper substrates using PDMS stamps derived from 3-D printed molds. ............................................................... 195  Figure 9.1 Smartphone-based imaging of fixed SK-BR3 cells labeled with F8BT/Dex Pdots in the absence (─) and the presence (+) of TAC-HER2. Scale bars are 500 microns. F8BT Pdot labeled cells appeared green yellow in the smartphone image. The images are in RGB. ....................... 201   xxxii  List of Abbreviations 1H NMR Proton nuclear magnetic resonance  2PE Two-photon excitation 3-D Three-dimensional a.u. Arbitrary units A555 Alexa Fluor 555 dye Abs. Absorbance AIDS Acquired immunodeficiency syndrome AP Alkaline phosphatase API 1-(3-Aminopropyl)imidazole API-Dex API-modified dextran API-MNP API-modified iron oxide nanoparticle APTES (3-Aminopropyl)triethoxysilane ATR Attenuated total reflectance Au NPs Gold nanoparticles Au@Pd Gold-palladium  BP Bandpass BPE 1,2-di(4-pyridyl)ethylene  BSA Bovine serum albumin BSA-FITC FITC-labeled bovine serum albumin  BTA benzene-tricarboxamide  CCD Charge-coupled device CE Capillary electrophoresis CNMEHPPV Poly[2-methoxy-5-(2-ethylhexyloxy)-1,4-(1-cyanovinylene-1,4-phenylene)] CN-PPV Cyano-poly(phenylene vinylene)  ConA Concanavalin A CPT Camptothecin CR Contrast ratio CTCs Circulating tumour cells CWL Centre wavelength DAB Diaminobenzidine DAPI 4',6-Diamidino-2-phenylindole xxxiii  Dex Dextran Dex-MNP Dextran-coated iron oxide nanoparticles Dex-QD Dextran-coated quantum dots DHLA Dihydrolipoic acid DIC Differential interference contrast DLS Dynamic light scattering DNA Deoxyribonucleic acid D-p-API Dextran pendantly modified with API D-p-DHLA Dextran pendantly modified with DHLA D-t-DHLA Dextran terminally modified with DHLA DTNB 5,5’-dithio-bis-[2-nitrobenzoic acid] EDC 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide EDX Energy-dispersive x-ray spectroscopy EGFR Epidermal growth factor receptor ELISA Enzyme-linked immunosorbent assay Em. Emission EpCAM Epithelial cell adhesion molecule EPO Erythropoietin ER Estrogen receptor ESI Electrospray ionization Ex. Excitation Ext. Extinction F8BT Poly(9,9-dioctylfluorene-alt-benzothiadiazole) FACS Fluorescence activated cell sorting FC Flow cytometry FCF Franck-Condon factor FDA Food and Drug Administration FISH Fluorescence in situ hybridization FITC Fluorescein isothiocyanate Fl. Fluorescence FLIM Fluorescence lifetime imaging microscopy FOV Field-of-view FPN Fluorescent polymeric nanoparticles xxxiv  FRET Förster resonance energy transfer FTIR Fourier transform infrared spectroscopy FWHM Full-width-at-half-maximum Glc Glucose GSH L-Glutathione hCG Human gonadotropin  HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid HER2 Human epidermal growth factor receptor 2 His Histidine HIV Human immunodeficiency viruses HMAT Hexamethylazatriangulene HMDA Hexamethylenediamine HOMO Highest occupied molecular orbital HRP Horseradish peroxidase IC50 Half maximal inhibitory concentration IgG Immunoglobulin IHC Immunohistochemistry IPS Triethoxy-3-(2-imidazolin-1-yl)propylsilane  IR Infrared KRH Krebs-ringer-HEPES LA Lipoic acid LED Light-emitting diode LFA Lateral flow assay LP Longpass  LSPR Local surface plasmon resonance LUMO Lowest unoccupied molecular orbital MAA Mercaptoacetic acid MeOH Methanol MNP Iron-oxide nanoparticles MNP@QD Quantum dot-coated iron oxide nanoparticles MO Methyl orange MP Megapixel xxxv  MPA Mercaptopropionic acid mTOR Mammalian target of rapamycin MUC1 Mucin 1 MW Molecular weight MWCO Molecular weight cut-off NFC Near field communication NIH National institute of health nm Nanometer NMs Nanomaterials Norm. Normalized NP Nanoparticle NTA Nanoparticle tracking analyzer ODA Oxadiazole Ox-Dex Oxidized dextran OX-MNP Oxidized dextran-coated iron oxide nanoparticles PBS Phosphate buffered saline PDMS Polydimethyl siloxane Pdots Polymer dots PEG Polyethylene glycol PFBT Poly(9,9-dioctylfluorene-alt-benzothiadiazole) PFO Polydioctylfluorene PFOCTS Trichloro-(1H,1H,2H,2H-perfluorooctyl)silane PFPV Poly[{9,9-dioctyl-2,7-divinylene-fluorenylene-alt-co-{2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylene}] PISA Polymerization-induced self-assembly PL Photoluminescence PLA Polylactic acid PMT Photomultiplier tube POC Point-of-care PON Point-of-need PPE Poly(phenylene ethynylene) PPF Polymers with pendant fluorophores xxxvi  pPFPA Poly(pentafluorophenyl) acrylate  PR Progesterone receptor PSMA Polystyrene-co-maleic anhydride PS-PEG-COOH Polystyrene graft ethylene oxide functionalized with carboxy Pt NPs Platinum nanoparticles QDs Quantum dots QD-SABER Quantum dot - signal amplification by exchange reaction QY Quantum yield RCF Relative centrifugal force RGB Red green blue RNA Ribonucleic acid SBR Signal-to-background ratio CMOS Complementary metal-oxide-semiconductor sCMOS Scientific complementary metal-oxide-semiconductor SCPNs Single chain polymer nanoparticle SD Standard deviation SELEX Systematic evolution of ligands by exponential enrichment SEM Scanning electron microscope SERS Surface enhanced Raman spectroscopy SiO2 NPs Silica nanoparticles SiO2@QD Silica nanoparticle-quantum dot supra-nanoparticles SIP Smartphone-based imaging platform SLUMO Second-lowest unoccupied molecular orbital SMCC Succinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate  SNR Signal-to-noise ratio SP π-conjugated semiconducting polymers  ssDNA Single-stranded deoxyribonucleic acid xxxvii  TAC Tetrameric antibody complex TBE Tris-borate-ethylenediaminetetraacetic acid TCEP Tris(2-carboxyethyl) phosphine hydrochloride TCSPC Time-correlated single-photon counting TEM Transmission electron microscope TEOS Tetraethylorthosilicate THF Tetrahydrofuran TMAH Tetramethylammonium hydroxide solution  TNBS 2,4,6-trinitrobenzenesulfonic acid  TOPO Trioctylphosphine oxide TTC 2,3,5-triphenyltetrazolium chloride UCNP Upconversion nanoparticle UPy 2-ureido-pyrimidinone  USB Universal serial bus UV Ultraviolet X-MNP Epichlorohydrin-crosslinked iron oxide nanoparticles XPS X-ray photoelectron spectroscopy ZW Zwitterionic    xxxviii  Acknowledgements Firstly, I would like to sincerely thank my PhD supervisor/mentor, Prof. Russ Algar. I have learned so much from you throughout these years and am greatly appreciative.   I would also like to thank my committee members: Prof. Dan Bizzotto, Prof. Jay Kizhakkedathu, and Prof. Roman Krems for their helpful feedback and guidance.  Much of the research within this thesis could not have been conducted without the help of numerous research staff within the UBC Chemistry Department and across the UBC Vancouver campus. Special thanks to: Ben Herring, Derrick Horne, Elena Polishchuk, Francis Manalastas, Gethin Owen, Jessie Chen, Pritesh Padhiar, and Saeid Kamal. I would like to extend this thanks to Prof. Zac Hudson and his talented research group for their collaboration in several research projects.  I have had the pleasure of working alongside many fantastic lab colleagues. I would like to give my dearest thanks to the Algar Lab members, for their collaborative efforts have been instrumental in the creation of this thesis. I will miss our random office chats, dominion games on the cardboard box, and potlucks.  Kelly Rees, you have been my greatest support, both inside and outside of the lab. Thank-you for always helping me see things in a more logical sense. You’ve kept me grounded whenever I came across any challenge.  To my parents, thank-you for all of the sacrifices you have made, I cannot imagine how difficult it must have been to settle into a new country to provide a more privileged life for your children, and for that I am grateful.    1   Introduction   1.1 Overview  This thesis presents research toward the development of luminescent materials for applications in bioanalysis and imaging, and particularly the immunofluorescent labeling of cells. Many different types of fluorescent nanoparticle materials are evaluated for this purpose, whether in the context of novel tools for biomedical research or toward point-of-care molecular diagnostics with smartphone-based devices. This introductory chapter provides the conceptual background and literature review necessary to appreciate the motivation, technical details, and scope and impact of this research.   1.2 Fluorescence  The spontaneous emission of ultraviolet, visible, or infrared light by an atom, molecule, or lattice is luminescence. If the light emitted from a material is the result of photon absorption, then the process is known as photoluminescence (PL). Other forms of luminescence rely on different sources of energy. These include chemiluminescence, bioluminescence, and electroluminescence, which rely on a chemical reaction, biochemical reaction, and an electrical current, respectively. Fluorescence is a type of PL that is the result of the transition of an electron from a higher energy excited state to a lower energy ground state, without a change in spin multiplicity. The polyatomic molecules that exhibit fluorescence are known as fluorophores. Fluorescence is widely used as a probe to gain physical insight and make analytical measurements of chemical and biological systems.  1.2.1 The Perrin-Jablonski diagram The electronic transitions that occur during the process of fluorescence are typically modelled using a Jablonski diagram. As illustrated in Figure 1.1, the different electronic states of a fluorophore are represented by a potential energy well, with multiple vibrational energy levels superimposed. Although the molecules do have rotational energy levels, they are not displayed 2  within the diagram for reasons of simplicity and the lesser relevance to condensed phases (i.e. molecules do not rotate freely in liquid or solids and so do not have well defined rotational energy levels).   In fluorescence, the fluorophore absorbs the energy of a photon (typically 200–800 nm wavelength). This energy promotes an electron from the lowest vibrational energy level (𝜈଴) of the highest occupied molecular orbital (HOMO, 𝑆଴) to any vibrational energy level (𝜈௡) of the lowest unoccupied molecular orbital (LUMO, 𝑆ଵ) or second-lowest unoccupied molecular orbital (SLUMO, Sଶ𝜈௡). The absorption process is fast, on the order of 10–15 s. (Excitation from the spontaneous absorption of two photons is also possible under certain conditions; however, this section will focus only on the excitation of fluorescence by one-photon absorption). The terms HOMO and LUMO are associated with the ground electronic state and excited electronic state respectively.   Once a molecule is excited and its LUMO occupied, vibrational relaxation occurs with loss of energy as heat. Kasha’s rule states that “the emitting level of a given multiplicity is the lowest excited level of that multiplicity” and is a consequence of vibrational relaxation being a faster process than an electronic transition to the ground state.1 Following vibrational relaxation to 𝑆ଵ𝜈଴, the electron transitions back down to any vibrational energy level within the ground state (𝑆଴𝜈௡), with the release of a photon to generate fluorescence. The energy of the photon matches the energy difference between the initial and final states. Exceptions to Kasha’s rule are rare, but have been observed for molecules such as azulene and cycl[3.3.3]azine. These molecules have a large energy gap between S1 and S2, allowing for fluorescence that corresponds to the S2 to S1 transition.2  Non-radiative relaxation processes compete with the radiative fluorescence process. These non-radiative relaxation processes include internal conversion from Sଵ𝜈଴ or Sଶ𝜈଴ to an isoenergetic energy vibrational level of the lower-energy electronic state, from which the electron undergoes vibrational relaxation again (see Figure 1.1B). Delayed fluorescence and phosphorescence (see Figure 1.1C) is another radiative relaxation pathway that is possible after a photon is absorbed. Both processes require intersystem crossing from 𝑆ଵ𝜈଴ to 𝑇ଵ𝜈௡, the latter being an isoenergetic level of the lowest-energy triplet state.  3    Figure 1.1. Jablonski diagrams illustrating the processes of photon absorption, fluorescence, and competing non-radiative processes. The y-axis is potential energy and the x-axis represents the nuclear coordinate. Singlet (𝑆௡) and triplet (𝑇௡) electronic states are shown as potential energy wells, each with vibrational states (𝜈௡) superimposed. (A) The processes involved in a fluorescence relaxation pathway: (i) photon absorption results in (for example) electron promotion from 𝑆଴𝜈଴ to 𝑆ଵ𝜈ଶ; (ii) the electron transitions to 𝑆ଵ𝜈଴ by vibrational relaxation; and (iii) radiative relaxation to the ground electronic state, generating a photon. (B) Processes involved in a non-radiative relaxation pathway. Photon absorption results in (for example) a transition from 𝑆଴𝜈଴ to (i) 𝑆ଵ𝜈ଶ or (ii) 𝑆ଶ𝜈ଶ. From 𝑆ଶ𝜈ଶ, vibrational relaxation transitions the electron to 𝑆ଶ𝜈଴, where (iv) internal conversion to 𝑆ଵ𝜈଻ occurs. Next, vibrational relaxation to 𝑆ଵ𝜈଴ occurs, after which (v) internal conversion to 𝑆଴𝜈ଵଷ is followed by vibrational relaxation to 𝑆଴𝜈଴. (C) Other relaxation pathways: (i) photon absorption excites an electron from 𝑆଴𝜈଴ to 𝑆ଵ𝜈ଶ, followed by (ii) vibrational relaxation to 𝑆ଵ𝜈଴. Next, (iii) intersystem crossing to 𝑇ଵ𝜈ଵ occurs and is followed by (ii) vibrational relaxation to 𝑇ଵ𝜈଴. From 𝑇ଵ𝜈଴, radiative relaxation can occur via phosphorescence (not shown), reverse intersystem crossing and delayed fluorescence (not shown), or by non-radiative pathways similar to B.   1.2.2 Photon absorption Molecular absorption of a photon causes an electronic transition from a ground state to an excited state (i.e. HOMO to LUMO or SLUMO). HOMO-LUMO transitions are typically n to π* or π to π*, where n, π, and π* represent nonbonding, pi-bonding, and pi-antibonding orbitals, respectively. The portion of a molecule that absorbs a photon is known as a chromophore. In pi-conjugated systems, the extent of conjugation approximately determines the wavelength (i.e. photon energy) 4  at which absorption will occur.3 A chromophore with a longer conjugated system absorbs longer-wavelength (i.e. lower-energy) photons, and vice versa.   When a molecule is illuminated with photons, the probability of photon absorption by the molecule is dependent on several selection rules:  (1) Resonance condition The energy of an incident photon must match the energy of the transition (∆𝐸) from the lowest vibrational energy level of the ground state, 𝐸(𝑆଴𝜈଴), to any vibrational energy level in the excited state, 𝐸(𝑆௡𝜈௠), as per in Eqn. (1.1), where h is Planck’s constant and f is the frequency of the photon.   (2) Photoselection principle Absorption of a photon by a molecule causes a shift in that molecule’s electron density, represented by a transition dipole, denoted by the vector, 𝜇. The photon absorbed by the molecule is ideally polarized to have an electric field vector (𝐸) that is aligned with the transition dipole. As described in Eqn. (1.2), the probability of photon absorption is proportional to the dot product of 𝐸 and 𝜇. The angle between these two vectors is 𝛼, and thus the maximal probability of photon absorption occurs when this angle is 0˚ (i.e. parallel). When the two vectors are perpendicular (i.e. 90 ˚), photon absorption does not occur.3    (3) Electronic selection rules A set of conditions must be satisfied for an electronic transition to be allowed.4 The transitions that are not allowed are known as forbidden transitions. These conditions correspond to the conservation of angular momentum within a molecule:   Total orbital angular momentum quantum number, Λ, satisfies ∆Λ =  0, +1, −1  𝐸௣௛௢௧௢௡  =  ℎ𝑓 =  ∆𝐸௧௥௔௡௦௜௧௜௢௡  =  𝐸(𝑆௡𝜈௠)  −  𝐸(𝑆଴𝜈଴) (1.1)  𝑃஺௕௦   ∝  (𝐸 ∙ 𝜇)ଶ  =  𝐸ଶ 𝜇ଶ 𝑐𝑜𝑠ଶ 𝛼 (1.2) 5   Total spin angular momentum quantum number, Σ, satisfies ΔΣ = 0   Total angular momentum quantum number, Ω (= Λ + Σ), satisfies ΔΩ =  0, +1, −1.   The total spin cannot change: ΔS =  0  1.2.2.1 Absorption spectrum Fluorescent molecules have absorption and emission spectra. The shape of these bands is largely defined by the vibrational energy levels, whereas the difference in electronic energy levels determines the spectral position (wavelength) of the bands. (In solution, the rotational energy levels tend to be smeared out in condensed phases due to solvent interactions). Under common conditions, the electronic transition associated with photon absorption begins at the lowest-energy vibrational level of the ground electronic state (𝑆଴𝜈଴). The relative occupancy of the vibrational states is determined by the Boltzmann distribution function, Eqn. (1.3):   𝑁୧𝑁 =  𝑒(ି∆ாೡ೔್/௞್்) (1.3)  The number of molecules (i.e. electrons) in a specific vibrational state (𝑖 =  0, 1, 2 …) is denoted by 𝑁୧. The energy difference between two vibrational states is ∆𝐸௩௜௕, T is the absolute temperature, and 𝑘௕ is the Boltzmann constant. The thermal energy is 𝑘௕𝑇, and, for a molecule at room temperature (~298 K), the relative population of higher vibrational energy levels within the ground state is negligible.  The shape and intensity of the absorption and emission bands is partially determined by the Franck-Condon Factor (FCF), which relates the probability of an electronic transition between an initial and final state to the relative band intensity through quantum mechanical formulations. The other selection rules described in section 1.2.2 also modulate the intensity of these bands (i.e. the probability of a transition). As shown in Eqn. (1.4), the initial and final vibrational energy levels can be defined as wavefunctions (𝜓), and the square of the overlap integral of these wavefunctions is the FCF. The larger the FCF, the greater the probability of an electronic transition, and thus greater band intensity.5  6   𝐹𝐶𝐹 =  ൤න 𝜓௙௜௡௔௟∗ ⋅ 𝜓௜௡௜௧௜௔௟𝑑𝜏൨௩௜௕ଶ (1.4)   Electronic transitions are confined by the Franck-Condon principle (see Figure 1.2), which states that all transitions on a Jablonski diagram are vertical (i.e. occur without a change in the position of the nuclei).3 This principle is a consequence of the slow nuclear motion (~10–12 s) relative to the timescale of the photon absorption (~10–15 s). Both the FCF and Frank-Condon principle apply equally well to fluorescence transitions.     Figure 1.2. An illustration of the Franck-Condon principle for electronic transitions from the ground electronic state (𝑆଴) to an excited state (𝑆ଵ). (A) The absorption/emission intensity of an electronic transition is proportional to the overlap between the initial and final vibronic wavefunctions. The blue line represents the most probable transition from 𝑆଴𝜈଴ to 𝑆ଵ𝜈ଶ, which has the greatest overlap between vibronic wavefunctions. The orange line represents the most probable transition from 𝑆ଵ𝜈଴ to 𝑆଴𝜈ଶ. The probability density functions are shown in the shaded areas of the wavefunctions. (B) Corresponding absorption and emission spectra and possible transitions from panel A. Note that the individual transitions are typically only observed in the gas phase or at very low temperatures. In solution, around room temperature, the spectral features are usually broadened into a smooth shape (shaded regions). As per the FCFs, the absorption and emission spectra are mirror images of each other. 7  The probability of an electronic transition is often represented using its molar absorption coefficient (𝜀, L mol ି1cm ି1) or its oscillator strength (f, unitless).3 The molar absorption coefficient is related to the equivalent cross-sectional area of absorption for a chromophore (𝜎, cm2), given in Eqn. (1.5):   𝜎 =  2.303 𝜀𝑐𝑛 =  2.303 𝜀𝑐(𝑁Ac/10ଷ)  =  3.82 × 10ିଶଵ 𝜀 (1.5)  where n is the number of molecules per cm3, c is the concentration of the chromophore, and NA is Avogadro’s number. A molar absorption coefficient can be measured at a specific frequency or wavelength. A larger molar absorption coefficient indicates a more probable transition and thus a larger band intensity for a given wavelength. According to the selection rules in section 1.2.2, transitions that are spin-allowed and orbital-allowed have 𝜀 ≈ 103 to 105. For electronic transitions that are spin-allowed, but orbital-forbidden, 𝜀 ≈ 100 to 103. The electronic transitions that are spin forbidden but orbital-allowed have molar absorption coefficients between 𝜀 ≈ 10–5 and 100.3  The fraction of light absorbed by a population of a given chromophore is described by the Beer-Lambert Law, given in Eqn. (1.6):   𝐴(𝜆)  =  −𝑙𝑜𝑔𝑇(𝜆)  =  𝑙𝑜𝑔 ቆ𝐼଴(𝜆)𝐼(𝜆)ቇ  =  𝜀(𝜆)𝑏𝑐 (1.6)  where T is the transmittance through a sample containing a chromophore at concentration, c. The incident light intensity is represented as 𝐼଴(𝜆), and the light intensity that reaches the detector is represented as 𝐼(𝜆). The amount of light that is transmitted through the sample is also dependent on the path length, b, of the sample cell.   1.2.3 Excited-state processes There are several different processes that can occur once an electron occupies an excited state. Generally, these processes are categorized into radiative (i.e. with light emission) or non-radiative (i.e. heat generation) relaxation pathways. A chromophore, prior to light absorption, is in its 8  equilibrium state, with an optimal nuclear configuration within its microenvironment. During the process of light absorption by the chromophore, a perturbation of the electron density within the chromophore occurs, leading to a nonequilibrium excited state with an unoptimized nuclear configuration. A chromophore in its nonequilibrium excited state undergoes vibrational relaxation, generating heat in the process.3 This non-radiative process is sufficiently efficient that relaxation occurs to the lowest vibrational energy level (equilibrium) of an excited state (𝑆௠𝜈଴, 𝑚 =  1, 2) prior to any other non-radiative or radiative relaxation mechanisms (see Figure 1.1A, ii).  Internal conversion is a non-radiative, isoenergetic (i.e. horizontal) transition of an electron from a higher electronic state to a lower electronic state (e.g. S2 to S1, S1 to S0) without a change in spin multiplicity (see Figure 1.1B, iv). This horizontal transition, which proceeds without the loss of energy (i.e. heat), is followed by vibrational relaxation and loss of energy as heat until the ground vibrational level of the new electronic state is reached. According to the Energy Gap Law, FCFs increase as the energy gap between two states decreases.2 In the vast majority of cases, internal conversion occurs on the timescale of 10–13 to 10–11 s for a transition between two electronic excited states (e.g. S2 to S1), but is slower (10–9 to 10–7 s) for a transition to the ground electronic state because of a larger energy gap for S1-S0 versus S2-S1. It is for that reason that fluorescence is almost invariably observed from S1.  Intersystem crossing is a non-radiative process that occurs with a change in spin-multiplicity (see Figure 1.1C, iii). This process occurs between isoenergetic states and is a horizontal transition from the excited singlet state (𝑆ଵ𝜈଴) to the triplet state (𝑇ଵ𝜈௡). Despite being a spin-forbidden process, there is a nonzero probability of this process occurring due to spin-orbit coupling (i.e. coupling between the spin and orbital magnetic moments). Intersystem crossing is followed by vibrational relaxation to 𝑇ଵ𝜈௢.  Radiative relaxation pathways include fluorescence and phosphorescence (see Figure 1.1). An electron transitions back to ground state from an excited singlet or triplet electronic state occurs for fluorescence and phosphorescence, respectively. Both of these radiative relaxation pathways occur from the equilibrium excited state, having first undergone relaxation down to the lowest vibrational energy level of the corresponding excited state (see Figure 1.1A, ii). The fluorescence 9  process is an electronic transition from 𝑆ଵ𝜈଴ to 𝑆଴𝜈୬ with spontaneous emission of a photon with a wavelength that matches the energy difference between the two states (see Figure 1.1A, iii). The timescale of fluorescence is typically on the order of 10–7 to 10–10 s, which corresponds to how long a molecule remains in the excited state prior to photon emission.3 The emission of a photon occurs on a timescale that is equivalent to photon absorption (10–15 s). The fluorescence emission spectrum of a molecule will generally be a mirror-image of its absorption spectra as a result of the Franck-Condon principle and FCFs. Emission spectra are also red-shifted in wavelength because of Kasha’s rule, FCFs, and solvent interactions (not discussed). The absolute difference in the absorbance and fluorescence emission peak wavelengths is the Stokes shift.   The phosphorescence process is an electronic transition from 𝑇ଵ𝜈଴ to 𝑆଴𝜈୬ with spontaneous emission of a photon with a wavelength that matches the energy difference between the two states. This process occurs with a change in spin-multiplicity, and although forbidden, occurs via spin-orbit coupling, analogous to intersystem crossing. Since it is a forbidden process, the timescale is relatively slow (10-6 to 10-3 s or longer), and therefore competitive non-radiative processes (e.g. intersystem crossing, vibrational relaxation) are more likely to occur with most molecules and under most conditions.   1.2.4 Quantum yield and fluorescence lifetime The quantum yield of fluorescence can be described as the ratio between excited-state molecules that relax via fluorescence and all photo-excited molecules (regardless of relaxation mechanism). More formally, the quantum yield is defined using the rates of fluorescence and non-radiative processes, Eqn. (1.7):   ௙   =𝑘௙𝑘௙ +  ∑ 𝑘௡௥ (1.7)  where ௙ is the quantum yield of a fluorophore, 𝑘௙ is the rate for fluorescence, and ∑ 𝑘௡௥  is the sum of the non-radiative rates, all of which are in units of s–1. As described in Eqn. (1.8), the non-radiative pathways include the rate of internal conversion, 𝑘௜௖ , rate of intersystem crossing, 𝑘௜௦௖ , rate of quenching, 𝑘௤ , rate of energy transfer, 𝑘௘௧ , and the rate of photodegradation, 𝑘௣ௗ . 10   These non-radiative pathways are in competition with fluorescence, and relaxation via these non-radiative pathways reduces the number of photons emitted. Values for the quantum yield of fluorescence range as 0 ≤ ௙ ≤  1.1  The depopulation of an excited state follows a first-order decay process defined by Eqn. (1.9):   where an excited state molecule is represented by 𝑆ଵ, t is the time following a pulsed excitation event, and 𝑘௙ and 𝑘௡௥ represent the fluorescence and non-radiative rates, respectively. The fluorescence photons that are emitted following pulsed excitation of a population of a fluorophore follow an exponential decay function, Eqn. (1.10):   where 𝐼(𝑡) is the fluorescence intensity at timepoint t, 𝐼଴ is the initial fluorescence intensity at t = 0, and 𝜏 is the fluorescence lifetime, defined by Eqn. (1.11):   The fluorescence lifetime is the characteristic amount of time that a molecule remains in its excited state, and is typically on the order of 10–10 to 10–7 s. Mathematically, the fluorescence lifetime corresponds to the time at which the fluorescence intensity has decayed to 37% of the value of 𝐼଴ (i.e. 63% of the excited state molecules have returned to the ground state).   ෍ 𝑘௡௥ =  𝑘௜௖ +  𝑘௜௦௖ + 𝑘௤ + 𝑘௘௧ + 𝑘௣ௗ + ⋯ (1.8)  −𝑑[𝑆ଵ(𝑡)]𝑑𝑡  = (𝑘௙ +  ෍ 𝑘௡௥)[𝑆ଵ(𝑡)] (1.9)  𝐼(𝑡)  =  𝐼଴𝑒𝑥𝑝(−𝑡/𝜏) (1.10)  τ  =1𝑘௙ +  ∑ 𝑘௡௥ =  ௙𝑘௙ (1.11) 11  1.2.5 Photodegradation of a fluorophore It well known that the continuous or repeated excitation of many fluorophores results in a gradual and unrecoverable decrease in fluorescence intensity.6 This irreversible photodegradation is often called “photobleaching” and is usually a result of permanent molecular changes (e.g. covalent bonds breaking, molecular rearrangements). These molecular changes prevent the fluorophore from recovering to its original ground state, thereby preventing further fluorescence. Several factors affect the rate of photobleaching, including the intensity and wavelength of the excitation light, temperature, and oxygen level.6 The susceptibility to photodegradation varies between dyes, for example, fluorescein is more susceptible to photodegradation than its spectrally analogous Alexa Fluor dye counterpart.  1.2.6 Factors affecting fluorescence The photophysical parameters of fluorophores (e.g. quantum yield, lifetime) are sensitive to the fluorophore’s local microenvironment. Examples of relevant properties of a microenvironment include temperature, viscosity, polarity, pH, and the presence of quenchers.3 A temperature increase typically results in a decrease in quantum yield and lifetime as a consequence of an increase in the net rate of non-radiative relaxation. In contrast, a change in microenvironment that slows down molecule motions, such as an increase in viscosity or decrease in temperature, will result in an increase in the quantum yield from a decrease in the net rate of non-radiative relaxation. Other solutes, called quenchers, can also introduce new non-radiative relaxation pathways that compete with fluorescence.  The polarity of the solvent in which a fluorophore is dissolved may affect the spectral position of fluorescence. Most commonly, a more polar solvent will lead to an increase in the Stokes shift because a more polar solvent is better at stabilizing the equilibrium excited state. Fluorophores that exhibit larger increases in polarity moving from their ground state to their excited state will be more sensitive to these solvent effects.5  The pH of a microenvironment can change the protonation of a fluorophore. These changes can affect the properties of the nominal fluorophore, including its emission band shape and spectral position, quantum yield, and lifetime. The reason is that a change in protonation can result in a 12  change in the extent of a molecule’s conjugation. For example, fluorescein is capable of existing as a cation, neutral, anion, or dianion, with pKa values for acidic protons of 2.1, 4.3, and 6.4. The neutral molecule has a peak molar absorptivity of 11 000 L cm–1 mol–1 at 434 nm, whereas the dianion has a peak molar absorptivity of 76 900 L cm–1 mol–1 at a red-shifted wavelength of 490 nm.3  1.2.7 Fluorescence measurements and imaging Fluorescence measurements are generally categorized as being either steady-state or time-resolved. Instruments measure fluorescence intensity or photon counts, and different technical configurations for measurements enable characterization of photophysical parameters such as fluorescence spectra, quantum yields, lifetimes, and anisotropies.   Steady-state fluorescence measurements are made with constant illumination of a sample by a light source, such that the excitation and relaxation rates of a fluorophore are at equilibrium. This type of measurement is common with a spectrofluorimeters and fluorescence microscopes, which consist of optical components and wavelength selectors that guide excitation light to the sample and emission light to a photodetector. The most common types of steady-state fluorescence measurements include:   non-spectral measurements, where fluorescence intensity is measured within a fixed emission wavelength range for a fixed excitation wavelength range;  emission spectra, where fluorescence intensity as a function of emission wavelength with a fixed excitation wavelength; and   excitation spectra, where fluorescence intensity is measured at a fixed emission wavelength as a function of excitation wavelength.   The fluorescence intensity, 𝐹(𝜆′), as measured by most instruments is approximated by Eqn. (1.12):    𝐹(𝜆′)  =  2.303௙𝑏𝑐 ඵ 𝑃଴(𝜆) 𝜀(𝜆) 𝐾(𝜆′) 𝐿(𝜆′) 𝑑𝜆 𝑑𝜆′ (1.12) 13  where b is the sample pathlength; c is the fluorophore concentration; 𝑃଴(𝜆) is the incident light spectral power density; 𝜀(𝜆) is the wavelength-dependent molar absorptivity of the fluorophore; 𝐾(𝜆′) is the wavelength-dependent light collection and detection efficiency of the instrument; and 𝐿(𝜆′) is the function specific for the fluorophore’s bandshape.7  In contrast to steady-state fluorescence measurements, the most common form of time-resolved measurement utilizes pulsed excitation, which allows for measuring fluorescence intensity as a function of time under nonequilibrium conditions. One technique for time-resolved measurements, which is a feature of many fluorescence microplate readers, is measurement of fluorescence intensity by integrating over a defined time period after a defined delay time following pulsed excitation. The timescales of the delay and integration are typically microseconds up to milliseconds. The delay can improve the signal-to-background ratio of measurements by preventing the collection of unwanted scattered light or fluorescence from sample components that have a much shorter lifetime than the analytical fluorescent probe. Another technique for time-resolved fluorescence measurements is time-correlated single-photon counting (TCSPC), which is the most popular method for measuring the fluorescence lifetimes. Photon counts are binned as a function of time after an excitation pulse to reproduce the decay in fluorescence intensity over timescales typically measured in picoseconds, nanoseconds, or microseconds.1  Fluorescence microscopy is family of techniques that acquire images of fluorescence properties. The simplest format is images that map the spatial distribution of fluorescence intensity.8 A sample is illuminated with a narrow band of excitation wavelengths to generate fluorescence at longer wavelengths. Aside from lenses, key optical components of a fluorescence microscope include a dichroic mirror, which reflects/transmits light shorter/longer than a specific wavelength cutoff, and an emission filter, which selects a bandwidth of light for detection by a camera. The most common cameras utilized in fluorescence microscopy are those based on charge-coupled device (CCD) and scientific complementary metal-oxide-semiconductor (sCMOS) image sensors. To a first approximation, these detectors consist of matrix of photodiode-like pixels, the signals from which are proportional to the incident fluorescence light intensity and ultimately converted into image pixel values. Specialized fluorescence microscopy systems can be designed to measure different fluorescent parameters (e.g. lifetime, anisotropy) in an imaging format. For example, a 14  microscopic imaging analog of TCSPC is fluorescence lifetime imaging (FLIM), which acquires a time-resolved fluorescence intensity decay curve at each pixel within an image.8  1.3 Immunohistochemistry  Immunohistochemistry (IHC) is a vast field of bioanalysis and bioimaging that, for diagnostic and prognostic purposes, aims to measure specific types of cell lines via specific labeling. The most common application of IHC is in the detection of prognostic markers for cancer; however, IHC is also used in the detection of pathogens such as Cytomegalo virus, Hepatitis B virus, Hepatitis C virus, pneumococci and other bacterial and protozoal pathogens embedded within tissues.9 IHC is also utilized in genetics, assessing neurodegenerative disorders, brain trauma, and the diagnosis of muscular dystrophy.9 The term immuno comes from the antibodies utilized in this technique, whereas the term histo comes from the tissue-sections/cells that are typically analyzed via this technique.10 This technique exploits the highly specific binding between an antibody and an antigen (e.g. protein) on or within a mammalian cell. The antibody typically recognizes and binds to the epitope region of the antigen.11 The results of this technique are typically semi-quantitative, indicating the presence of an antigen utilizing either chromogenic or fluorescent reporters. Cells and tissues are typically labeled in a primary (direct) or secondary (indirect) fashion (see Figure 1.3).11 With primary labeling, the antibody that recognizes the antigen is also conjugated with the reporter. In contrast, the indirect method utilizes a secondary antibody, which is conjugated to the reporter and binds specifically to the antigen-bound primary antibody. The indirect method is feasible because antibodies made in one host animal (e.g. mouse, rat, rabbit, goat) can be made to recognize antibodies made in another host animal.  Another method for cellular labeling is the utilization of tetrameric antibody complexes (TACs). TACs are antigen targeting complexes that are comprised of four antibodies, where two antibodies are directed towards two separate antigens, and the other two antibodies hold the former two antibodies together. An early example, demonstrated by Wognum et al. in 1987, utilized dye-labeled TACs as a stain for cells in flow cytometry measurements.12 Currently, TACs are commonly used with dextran-coated magnetic nanoparticles (See Section 1.4.3), where one of the antibodies is specific for dextran and the other antibody is specific for a cell type of interest (e.g. 15  monocytes, T cells, B cells)13. Later in this thesis, TACs will be used with luminescent nanoparticles.   A common application of immunohistochemistry is antigen profiling. Antigen profiling (also known as cell phenotyping, cell profiling, and molecular profiling) aims to quantify antigen expression patterns associated with different cell types with the goal of providing a diagnostic antigenic “fingerprint”.     Figure 1.3 Primary (direct) versus secondary (indirect) immunolabeling strategies. A nanoparticle (NP) label is conjugated with either a primary or secondary antibody. The primary antibody binds specifically to the cell antigen, whereas the secondary antibody binds to a primary antibody.   1.3.1 Chromogenic immunohistochemistry  When a chromogenic reporter is used, an enzyme such as alkaline phosphatase (AP) or horseradish peroxidase (HRP) is linked to a primary or secondary antibody.11 Enzymatic turnover of the chromogenic substrate produces a visible colour. For example, the oxidation of 3,3′-diaminobenzidine (DAB) by hydrogen peroxide is catalyzed by HRP and forms a brown precipitate that can be visualized under a light microscope. The advantages of utilizing a chromogenic reporter for IHC includes greater sensitivity for antigen quantification from the enzymatic amplification. Disadvantages of this method include poorly feasible multiplexed profiling of multiple antigens and lower resolution. 16  1.3.2 Immunofluorescence Immunofluorescence is an IHC technique that is typically used in conjunction with fluorescence microscopy to analyze labeled cells/tissues.10,14 Primary or secondary antibodies are labeled with fluorescent dyes or fluorescent proteins, which serve as reporters. Fluorescent immunolabels can also be utilized in conjunction with cellular stains (e.g. 4',6-Diamidino-2-phenylindole, DAPI, which stains DNA within the nucleus). Imaging of immunolabeled cells on a fluorescence microscope is useful for evaluating antigenic expression profiles of various cell types (e.g. leukocytes, erythrocytes, cancer cells) and tissues (e.g. tumor slices). The fluorescence microscope is equipped with specific optics (e.g. filters, dichroic mirrors, lasers) that are tailored to the fluorophores in use, where different colours of fluorescence are associated with expression of different antigens. A significant advantage of immunofluorescence over chromogenic immunostaining is the capability to analyze multiple antigens simultaneously (i.e. multiplexing). The wide commercial availability of fluorescent dyes allows for pairing of different colours of dyes with the different antigens to be analyzed. Furthermore, colocalization studies are feasible with immunofluorescence, allowing for analysis of multiple protein targets at a specific cell or tissue location. Challenges from photobleaching are a disadvantage of immunofluorescence.   1.3.3 Flow cytometry Flow cytometry (FC) provides information about antigen expression on cells labeled with fluorescent-antibody probes. Different cellular antigens are labeled with dyes of different colours. The cells, dispersed within a liquid stream, are passed through a nozzle that forces and aligns the cells in single file. The cells then pass through an excitation source (usually a laser beam) and the fluorescence is separated by wavelength using a combination of dichroic mirrors, emission filters, and photodetectors (usually photomultiplier tubes, PMTs). The intensity of the fluorescence signals is correlated to the level of antigen expression on the cell. The FC system also measures the light scattered off the cells (e.g. forward scattered, side scattered) to obtain information about cell size and morphology, helping to verify that a cell is being detected rather than debris.15 When a homogeneous population of cells is desired, the most widely utilized technique is fluorescence-activated-cell-sorting (FACS), which is a type of flow cytometry. In FACS, heterogeneous populations of cells are labeled, typically with fluorescent antibodies, which target cell-specific antigens within the cell (i.e. intracellular) or on the cell surface (i.e. extracellular).16 FACS differs 17  from flow cytometry, as an electrical charging ring, which is attached to the nozzle, applies an opposite charge to the fluorescently labeled cells just prior to fluorescence intensity detection. Cells that are marked with the opposite charge are separated from the stream of cells through a series of charged deflection plates, which direct the cells towards separate test tubes.  1.4 General overview of nanoparticles  This section provides a general overview of the types of nanoparticles that are used in the subsequent chapters of this thesis. These nanoparticles include hard materials such as silica nanoparticles (SiO2 NPs), quantum dots (QDs), iron-oxide nanoparticles (MNPs), gold nanoparticles (Au NPs), and platinum nanoparticles (Pt NPs), as well as soft materials such as polymer dots (Pdots) and single-chain polymer nanoparticles (SCPNs). The level of detail provided for each material is commensurate with their prevalence within this thesis. When relevant, this section will review the unique physical and photophysical properties, bioconjugation strategies, and applications of each nanoparticle material in the context of bioanalysis and bioimaging.   1.4.1 Silica nanoparticles SiO2 NPs can be synthesized in a variety of shapes, including spherical, rod, ellipsoid, platelets, and sheets—all with at least one dimension less than 1 µm.17 These particles can also be prepared in several different forms, including amorphous, mesoporous, and hollow. The most popular synthesis methods are sol-gel processes, which can produce homogeneous particles under relatively mild conditions. These methods involve the hydrolysis and condensation of metal alkoxides (e.g. tetraethylorthosilicate, TEOS) under acidic or basic conditions. The formation of the nanoparticles begins with the nucleation of the clusters of hydrolyzed and condensed alkoxides, followed by growth. SiO2 NPs are characterized by their high specific surface area, with the number of silanol groups (−SiOH) scaling with the size of the particles. The silanol groups can act as chemical handles for the functionalization of these particles with silane molecules terminated with reactive functional groups (e.g. amine, carboxyl, epoxide) for anchoring with luminescent materials such as organic dyes,18 QDs,19 polymers (e.g. PEG), drugs, and targeting groups (e.g. antibodies, peptides).20 18  SiO2 NPs, unlike other materials in the nanometer size regime, do not acquire unique optical or electronic properties from their submicron size.20 Instead, it is the physical attributes of these materials that garner interest. These attributes include tunable properties such as size, morphology, porosity, and specific surface area. As such, SiO2 NPs are typically utilized as carriers for other functional materials such as contrast agents (e.g. fluorophores, paramagnetic complexes, radioisotopes) and therapeutics (e.g. anti-cancer drugs,21 anti-vascular drugs,22 nucleic acids23). For example, Hu et al. demonstrated the use of StreptAvidin-functionalized and FITC-coupled silica NPs for the specific labeling of hepatoma cells pre-labeled with biotin-TLS11a aptamer.24 Lin et al. developed multimodal SiO2 NPs that were doped with fluorescent Ru(bpy)32+ and further modified with Gd3+ chelates. These multimodal probes were used to label monocytes, utilizing both fluorescence and magnetic properties for optical and magnetic resonance imaging contrast. 25  1.4.2 Quantum dots QDs are fluorescent semiconductor nanocrystals that are approximately spherical in shape and 1–10 nm in diameter.26–29 These colloidal particles are composed of hundreds to thousands of atoms arranged in a crystalline lattice (Figure 1.4A). A wide variety of semiconductor materials are used to synthesize QDs, including II-VI materials (e.g. CdSe, CdTe), III-V materials (e.g. InP, InAs), IV-VI materials (e.g. PbS, PbSe), and group IV materials (e.g. Si).30,31 The nanometer size of QDs leads to unique optical properties that are remarkably different from the corresponding bulk semiconducting material. In particular, semiconductor QDs are brightly fluorescent whereas their bulk analogues are not.   When a semiconductor material absorbs a photon of light, an electron is excited across the band gap to form an electron-hole pair called an exciton. In the bulk size regime, the energy of the band gap is independent of the size of the semiconductor crystal; however, QDs are smaller than the Bohr exciton radius of the bulk semiconductor, leading to a phenomenon called quantum confinement.32 With quantum confinement, the band gap energy becomes dependent on the size of the nanocrystal, increasing as the size of the QD decreases (Figure 1.4B-C). Other consequences of quantum confinement include much stronger absorption of light and a significant fluorescence quantum yield, where the wavelengths (i.e. energies) of absorption and fluorescence emission can be tuned by the size of the QD. To date, the most widely utilized QDs are CdSe/ZnS 19  core/shell nanocrystals, where the CdSe core determines the optical properties and the ZnS shell protects and enhances those properties. CdSe/ZnS QDs have been widely utilized because of their well-established methods of synthesis and outstanding optical properties. Driven in part by the desire to develop materials without Cd, other QD materials are in development (e.g. InP, Si).33–35 The composition of the QD material is not just important because of the benefits or liabilities of the constituent elements, but also because the composition of the QD determines the range over which its fluorescence can be size-tuned (Figure 1.4D). It is also possible to tune the fluorescence of alloyed QDs (e.g. CdSexS1–x) by changing the composition without a change in nanocrystal size. Despite significant progress, the properties of most non-Cd QD materials do not yet match those of Cd-based QDs. The following text is largely written in the context of CdSe and related materials (e.g. CdTe, CdSeS) as benchmark QDs. These materials remain popular because of their highest-quality optical properties, established methods for synthesis, their fluorescence across the visible spectrum, and their commercial availability.36 Some alternative QD materials are now also available commercially (e.g. InP).   20   Figure 1.4. Overview of QDs. (A) (i) Atomistic illustration of two sizes of QD nanocrystal. (ii) High-resolution TEM image of a QD. Reprinted with permission from ref.28 (B) Absorption and fluorescence spectra for various sizes of QDs. Reprinted with permission from ref.37 (C) Size-tunable PL of CdSe QDs. The photograph was taken under UV illumination (365 nm). Reprinted with permission from ref. 28 (D) Approximate wavelength ranges over which the fluorescence of various QD materials can be tuned through control of nanocrystal size. The visible spectrum is between ca. 400–650 nm. Reprinted with permission from ref. 29  1.4.2.1 QDs in bioanalysis and imaging QDs exhibit both optical and physical properties that are highly advantageous for bioanalysis and imaging.36,38,39 Optically, QDs have broad absorption spectra that begin at the band gap energy and extend into the ultraviolet region, such that visible QD fluorescence can be excited over a wide range of wavelengths. In addition to being size-tunable, the fluorescence emission spectra of QDs are symmetric and narrow with full-width-at-half-maxima (FWHM) of 25–35 nm for relatively monodisperse samples. Taken together, these properties allow (i) the wavelength of QD fluorescence to be optimized and matched to the needs of an application, (ii) the fluorescence from multiple colours of QDs to be excited simultaneously, and (iii) each colour of fluorescence to be measured with minimal interference from other colours (i.e. minimal crosstalk). This multiplexing capability is very attractive in applications where multiple target biomarkers need to be analyzed simultaneously, as each target can be associated with a specific colour of QD.40 Importantly, QDs have quantum yields that are comparable to those of conventional organic 21  fluorescent dyes, and also have molar extinction coefficients (104–107 M–1 cm–1) that are typically 10–100-fold larger than those for organic dyes or fluorescent proteins (104–105 M–1 cm–1).26  These optical features make QDs very brightly fluorescent and thus potentially ideal for detection of small amounts of biomarkers or other analytes. Naturally, direct comparisons between QDs and fluorescent dyes and proteins have been made. For example, Petryayeva et al. demonstrated that QDs outperformed a bright and popular fluorescent dye (fluorescein) and fluorescent protein (R-phycoerythrin), exhibiting much higher brightness.41 However, an advantage fluorescent proteins that is unmatched by is the ability to genetically encode fluorescent proteins, often as fusion constructs with other proteins, inside live cells.42 This feature makes fluorescent proteins a powerful tool for intracellular sensing and imaging of endogenous proteins. QDs also exhibit superior resistance to photobleaching in comparison to dyes and fluorescent proteins, which is advantageous for signal integration, kinetic measurements, long-term monitoring, and the use of high-power excitation to maximize fluorescence signals.   Another advantage of using QDs for bioanalysis is the ability to functionalize their surfaces with small molecules, polymers, and biomolecules such as antibodies, enzymes, other proteins, nucleic acids (e.g. DNA, RNA, ssDNA, aptamers), lipids, and peptides.43 In many cases, the surface area of a QD is sufficiently large to conjugate multiple copies of these biomolecules, the role of which is most often targeted binding to specific analytes of interest. However, the functionalized surface of a QD may perturb the biological activity of the conjugated biomolecules. From this standpoint, conventional fluorescent dyes, which are much smaller and can be synthesized with a mono-reactive chemical moiety, are better suited to precise and minimally perturbative bioconjugation. As will be discussed further below, QDs are typically coated with small molecule ligands or polymers that determine their physical properties as a colloid, including stability with respect to pH and ionic strength, as well as biocompatibility.   1.4.2.2 Synthesis and functionalization strategies The optical properties of QDs are tailored through both the size of the nanocrystal and the type of semiconductor material. To maximize their advantages in bioanalysis, QDs should be synthesized in a controlled manner to minimize nanocrystal defects, obtain high brightness, and achieve monodispersity. The first documented synthesis of colloidal QDs was reported by Brus 22  and colleagues in 1984.44 They synthesized various sizes of core-only CdS QDs stabilised with a styrene/maleic anhydride co-polymer in water. Since then, the standard method for synthesizing QDs has become significantly different. The synthesis of high-quality, monodisperse and brightly fluorescent QDs became routine when researchers began injecting organometallic precursors of Cd and a chalcogen (i.e. S, Se or Te) into hot coordinating solvents such as tri-octylphosphine oxide (TOPO) under an inert atmosphere.45 Other key advances were the synthesis of core/shell QDs, where the core is epitaxially coated with a shell of a structurally compatible and higher band gap energy semiconductor (e.g. CdSe/ZnS),46 and also the use of safer and less volatile precursors.47 Another important advance was the realization that impurities in technical grade solvents were active in the synthesis of QDs but difficult to reproduce.48 Synthesis methods have thus been adapted to use solvents (e.g. octadecene) free of such impurities with controlled addition of coordinating ligands (e.g. long-chain alkyl acids; long-chain primary, secondary, or tertiary alkyl amines) for growth of QDs. Several reviews provide more details on the intricacies of QD synthesis.49–53  QDs can also be synthesized in water, which may be an attractive strategy given that most applications in bioanalysis will require QDs with aqueous compatibility; however, the overall quality of these QDs is significantly lower than QDs synthesized via high temperature organometallic methods. In particular, these aqueous-synthesized QDs often have poor quantum yields and large size polydispersity, resulting in lower brightness and spectrally broader PL.54   High-quality QDs synthesized by high-temperature organometallic methods are stabilised by hydrophobic ligands and are not suitable for most applications in bioanalysis until rendered hydrophilic. The two most prevalent methods for generating hydrophilic QDs are (i) exchange of the native hydrophobic ligands with hydrophilic ligands via mass action, and (ii) encapsulation with amphiphilic polymers or phospholipids.55–57 The hydrophilic ligands or polymers serve two purposes: maintaining a stable colloidal dispersion in an aqueous environment and providing functional groups for the conjugation of biomolecules to the QD.  1.4.2.3 Bioconjugation strategies Two general strategies are commonly used for the conjugation of biomolecules to QDs: direct 23  covalent coupling to the ligand or polymer coating on the QD; and non-covalent adsorption or metal-affinity coordination. The covalent coupling strategy forms new chemical bonds between functional groups on a biomolecule (e.g. amines, thiols, carboxyls) and functional groups on the QD coating. This strategy requires activating agents such as carbodiimides (e.g. 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide, EDC; couples amine and carboxyl groups) with  N-hydroxysuccinimide (NHS; increases the efficiency of the reaction), reactive groups such as maleimides (react with thiols) and succinimidyl esters (react with amines), or heterobifunctional crosslinkers such as succinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate (SMCC; links amine and thiol groups). The conjugation of biomolecules through adsorption often relies on electrostatic interactions between the biomolecule and the QD coating, whereas metal-affinity coordination relies on strong dative interactions between metal atoms on the inorganic surface of the QD and thiol-terminated linkers or polyhistidine tags on a biomolecule. Another approach to bioconjugation is to use either of the two strategies above to conjugate a QD with StreptAvidin, which permits simple attachment of biotinylated biomolecules. Additional and more detailed information about the bioconjugation of QDs can be found in several published reviews.27,43,58  Many types of biomolecules may be conjugated to QDs; however, antibodies and aptamers remain among the most common for bioanalysis and bioimaging. The role of both classes of biomolecules is to bind to an analyte of interest with high affinity and specificity. Aptamers are advantageous in comparison to antibodies because they can be selected for in vitro and chemically synthesized, resulting in lower cost and less batch-to-batch variation. Aptamers are also more robust than antibodies, smaller, and easier to design.59,60 Although some aptamers exhibit binding affinity and specificity that is comparable to an antibody, there are also many aptamers that have lower affinity for their targets than antibodies, and there are a multitude of targets for which aptamers have not yet been selected. Other reasons why antibodies are still utilized in conjunction with nanoparticles for both therapeutic and diagnostic applications include longer circulation times for in vivo studies, a well-developed market infrastructure, abundant support from finance and education, and rapid and sustained increase in the drug market share.61   Figure 1.5 and Figure 1.6 illustrate common conjugation strategies for antibodies and aptamers, respectively. The conjugation of aptamers to QDs relies on linkers attached to either the 5´ or 3´ 24  termini of the aptamer. These linkers tend to have terminal amine or thiol groups for crosslinking with functional groups (e.g. amine, carboxyl) associated with the ligand or polymer coating on the QD, or sometimes thiol groups can coordinate to the inorganic shell of the QD. Antibodies have many amine and carboxyl groups available for crosslinking, as well as disulphide bridges that can be reduced to yield free thiols for crosslinking.62 Alternatively, aptamers and antibodies can both be biotinylated to bind to QDs modified with StreptAvidin. Beyond crosslinking, a particularly effective and versatile approach to bioconjugation with QDs has been the high-affinity binding of polyhistidine-tagged proteins, antibodies, peptides, and oligonucleotides to the inorganic surface of QDs. These and other bioconjugation strategies have been reviewed in detail.27,58,63  25   Figure 1.5 Common bioconjugation strategies with peptides, proteins, and antibodies. (a) Coupling between thiol and amine groups using SMCC. (b) Coupling between carboxyl and amine groups using EDC activation (and often NHS, not shown). (c) Hydrazone coupling between hydrazide and aldehyde groups. (d) Binding of polyhistidine-tagged peptides (and proteins, not shown) to the inorganic shell of ligand-coated QDs. QDs can also be functionalized with coatings that display Ni2+-nitrilotriacetic acid groups for binding polyhistidine-tagged peptides (not shown). (e) Binding between StreptAvidin-QD conjugates and biotinylated antibodies. Figure adapted with permission from ref. 62. Copyright 2007 Nature Publishing Group. 26    Figure 1.6 Common bioconjugation strategies with aptamers and other oligonucleotides. (A) Coupling between carboxyl and amine groups using EDC activation (and often NHS, not shown). (B) Coupling between thiol and amine groups using SMCC. (C) Binding of a dithiol-terminated aptamer to the inorganic shell of ligand-coated QDs. (D) Binding of a biotinylated aptamer to StreptAvidin-QD conjugates. (E) Binding of polyhistidine-tagged aptamer to the inorganic shell of ligand-coated QDs. Parts of the figure have been adapted with permission from ref. 62. Copyright 2007 Nature Publishing Group.   27  1.4.2.4 Immunolabeling with QDs The many properties of QDs that make them desirable as a fluorescent reporter have motivated their development for immunofluorescence applications such as antigen profiling.62,64 Similar to fluorescent dyes, QDs must be conjugated to an antibody for use in antigen profiling experiments. Examples of common methods for immunolabeling cells with QDs are listed below:   Cells are directly labeled with primary antibody that has been conjugated with a QD through covalent chemistry;65,66  Cells are first labeled with a biotinylated primary antibody and then indirectly labeled with a QD-StreptAvidin conjugate;   Cells are directly labeled with primary antibody that has been conjugated with a QD through binding with Protein A or Protein G that was first conjugated to the QD;65,66  Cells are first labeled with a primary antibody and then indirectly labeled with a QD-conjugated secondary antibody.   Covalent strategies for conjugating antibodies and QDs are described in Section 1.4.2.3. These methods are simple and keep the antibodies intact but do not guarantee that the antibodies will be in the optimal orientation for binding with antigen. It must also be optimized to avoid aggregation, as antibodies can crosslink multiple QDs together. Aggregation is reduced by utilizing a heterobifunctional linker, such as SMCC, which has an amine-reactive (succinimidyl ester) end and a thiol-reactive (maleimide) end.62 The disulfide linkages that hold the two halves of an antibody together are reduced to thiol groups, which are then linked to the QD coating. This method ensures that antibody is conjugated in an orientation amenable to antigen binding. Less commonly, the carbohydrate residues on antibodies are oxidized to yield aldehyde groups suitable for reductive amination with hydrazide functionalized QDs.62   Noncovalent strategies for conjugating antibodies and QDs frequently utilize an adapter protein that enables spontaneous binding of the antibodies to the QDs.62,65,66 The most used adapter proteins are Protein A, Protein G, and the Avidin family of proteins (e.g. StreptAvidin, NeutrAvidin, Avidin). The adapter proteins may be conjugated to the QDs either covalently (using the methods noted in Section 1.4.2.3 ) or noncovalently. These proteins are commercially available 28  and can be obtained with a polyhistidine tag at a terminus. The tags can spontaneously and strongly bind to a QD and are a popular strategy for bioconjugates. Protein A/G are immunoglobulin binding proteins that target the crystalline region of the antibody (i.e. non-antigen-binding domain). Antibodies that are self-assembled to QD-Protein A/G conjugates are thus positioned in a favourable orientation for antigen binding. Both QD-StreptAvidin conjugates and biotinylated antibodies are commercially available. Although the StreptAvidin-biotin binding is strong and spontaneous, the biotin labels can be distributed over the entirety of the antibody, and thus a favourable orientation for antigen binding may not be achieved upon binding to QD-StreptAvidin conjugates.  A summary table is provided below, which outlines several different applications of QDs as immunolabels (Table 1.1). In 2007, Yezhelyev et al. demonstrated the first example of using QDs for multiplexed antigen profiling of various samples of breast cancers from the clinic.62 They prepared antibody conjugates of QDs with five different peak wavelengths (525, 565, 605, 655, and 705 nm) to target five different antigens: human epidermal receptor 2 (HER2), estrogen receptor (ER), progesterone receptor (PR), epidermal growth factor receptor (EGFR), and the mammalian target of rapamycin (mTOR). They were able to characterize the expression of these antigens in situ and obtained results comparable to traditional methods such as chromogenic IHC, FISH, and Western blotting. More recently, Zhou et al. demonstrated quantitative multiplexed antigen profiling of intracellular targets (HSP90, Ki-67, lamin A, calnexin, and β-tubulin) in HeLa cells with a 2.5-fold signal enhancement in comparison to cells labeled with single QD-conjugated antibodies.67 Their method, known as QD-SABER, combined QD and DNA nanotechnology for enhanced sensitivity, associating multiple QD labels with a single antibody. Other notable targets for QD immunolabeling include extracellular vesicles and pathogens (see Table 1.1).   29  Table 1.1 Summary table for applications of QD immunolabeling. Reference Antigen(s) Target QD Bioconjugate  Chemistry Yezhelyev et al, 2007 ref.62 HER2, ER, PR, EGFR, mTOR Cancer cell (MCF-7, BT474, MDA-231) QD-Primary antibody (SMCC coupling) Zrazhevskiy et al, 2013 ref.68 HSP90, Ki-67, lamin A, β-tubulin, Cox-4 Cancer cell (HeLa) QD-Protein A (NHS-ester coupling) Zhou et al, 2020  ref.67 HSP90, Ki-67, lamin A, β-tubulin Cancer cell (HeLa) QD-SABER Rodrigues et al 2019 ref.69 EpCAM, CD81, EphA2 Extracellular vesicles QD-StreptAvidin Hahn et al, 2005 ref.70 O157, H7 Pathogen (E. coli O157:H7) QD-StreptAvidin Lee-Montiel et al, 2015 ref.71 VEGFR, NRP Endothelial cells (HUVEC) QD-Primary antibody (Azide-alkyne) Min et al, 2015 ref.72 HA Avian Influenza A (H7N9) QD-Primary antibody (SMCC coupling) Chandan et al, 2016 ref.73 TNase Staphylococcus aureus QD-Primary antibody (NHS-ester coupling) Han et al, 2020 ref.66 αEGFR Cancer cell (A431) QD-Protein A  (His6 tagged) Hanifeh et al, 2019 ref.74  CD20 Diffuse large B-cell QD-Primary antibody (Azide-alkyne) 30  1.4.3 Magnetic nanoparticles Magnetic nanoparticles (MNPs) are spherical nanocrystals that vary in diameter between 5–30 nm. At this size regime, the magnetic properties are vastly different versus the bulk material.75 In a bulk iron oxide magnet, the magnetic moments are aligned without an external magnetic field. Above the Curie temperature for the bulk magnetic material, the magnetic moments become disordered and these materials lose their magnetization. Typically, the Curie temperature is ~850 K for bulk iron oxide magnetic materials. However, in MNPs, the disordering of magnetic moments occurs at ambient temperature due to the small volume that the NP occupies. When an external magnetic field is applied to the MNPs, the magnetic moments are then capable of aligning, thus returning the magnetism to the MNPs. This behaviour allows the magnetism of the MNPs to be turned on or off depending on the presence of an external magnetic field. The MNPs will aggregate when an external magnetic field is applied. Once the external magnetic field is removed, the MNPs can then redisperse into a colloidal solution. The MNPs that are used in molecular diagnostic applications are typically coated with an organic material that enables the conjugation of biological components to its surface. Common organic coatings on MNPs include dextran, polyethylene glycol (PEG), fatty acids, polypeptides, and polyacrylic acid.75 The magnetic properties of the MNPs have been used in applications ranging from contrast agents for magnetic resonance imaging76 to magnets for cell isolation.77,78   1.4.4 Metal nanoparticles Gold nanoparticles (Au NPs) can be prepared in a variety of shapes, such as spheres, rods, diamond-like,79 plates,80 branched, 80 nanocage,81 quasi-spherical,81 and several others. The shape of the NPs is dependent on the synthesis method. Au NPs can be synthesized in both organic solvents and aqueous solutions. Au NPs are typically stabilized in organic solvents via surfactants, which comprise hydrophobic alkyl chains. Conversely, aqueous stabilized Au NPs are typically coated with hydrophilic, small molecule ligands that have at least one thiol group for metal-ligand coordination to the gold nanoparticle surface. Once stabilized, these surfactants or ligands can be displaced with better stabilizing counterparts, such as PEG-based polymers. In the context of bioanalysis, a biomolecule can either be attached directly to the inorganic gold surface, via thiol-metal interactions, or attached covalently to the carboxy groups on the ligands that stabilize the Au NPs.82 The latter approach utilizes chemistries such as EDC/NHS. Incident light can be both 31  strongly scattered and absorbed by Au NPs. When light is absorbed, localized surface plasmons are generated at the interface between the Au NP surface and the surrounding dielectric medium. Surface plasmons are electromagnetic oscillations that travel in parallel to the interface between the surface and the medium.79 As the local refractive index changes, there is a shift in the plasmon absorption band. This band also shifts to longer wavelengths as the dimension(s) of the Au NP increases. A colloidal solution of larger spherical Au NPs (> 100 nm diameter) or aggregates of smaller spherical Au NPs appear as a blue/purple, whereas colloidal solutions of smaller spherical Au NPs (< 100 nm diameter) appears red. Au NPs are commonly used in nucleic acid assays, where hybridization between nucleic acid probes and targets assembles or disassembles clusters of Au NPs and changes their apparent colour.83 The strong scattering of Au NPs is also leveraged as a colorimetric label for lateral flow assays (LFAs).84 In LFAs, Au NPs are commonly attached to analyte-specific biomolecules such as nucleic acids or antibodies. Upon recognition of analyte by the Au NPs, a strong color change is visualized within the LFA strip, owing to the light scattering by the Au NPs. Another widespread application of Au NPs is as labels in surface enhanced Raman spectroscopy (SERS) experiments.82  Platinum nanoparticles (Pt NPs), much like Au NPs can also be prepared in a variety of shapes (e.g. sphere cube, tetrahedron, octahedron). The dimensions of these NPs can range from 1 nm up to hundreds of nanometers, depending on the type of shape. During synthesis, the Pt NPs are typically capped with a ligand which stabilizes the Pt NP in solution. For example, alkane thiols provide stability in organic solvents, whilst thiols bearing a polar group provide stability in hydrophilic environments. Similar to Au NPs, a common capping agent used with Pt NPs is sodium citrate and is widely regarded as a more environmentally-friendly alternative to other ligands.85,86 Other hydrophilic capping ligands include dihydrolipoic acid (DHLA), and glutathione (GSH), among several others. Like Au NPs, the optical properties of Pt NPs are typically dominated by light scattering and localized surface plasmon resonances. Pt NPs exhibit plasmon resonances from the UV (~215 nm) to blue (~450 nm) spectral range, scaling with the size of the NPs.87 Biological applications of Pt NPs have included the detection of alkaline phosphatase,88 as an anti-bacterial agent,89 folic acid-mediated cancer cell targeting,90 and anti-tumor activity,91 among many others.  32  1.4.5 Fluorescent polymeric nanoparticles Fluorescent polymeric nanoparticles (FPNs) are a broad class of labels that are synthesized using a wide variety of routes. FPNs are soft materials and, via the diversity of polymer chemistry, may provide for good control over emission brightness and color(s), nanoparticle size and surface chemistry, and other properties. Types of FPNs include (i) direct assembly of small π-conjugated organic dye oligomers into a single nanoparticle; (ii) dyes encapsulated within lipids or polymer particles; and (iii) nanoparticles comprised of fluorescent polymers.92 This thesis will focus on NPs that are prepared using intrinsically organic fluorescent polymers.   Two commonly utilized materials for the preparation of FPNs include polymers with pendant fluorophores (PPF) and π-conjugated semiconducting polymers (SP). PPFs are fluorescent polymers that have a non-conjugated backbone, with fluorescent emissive groups attached as sidechains. An advantage of PPFs is the synthetic flexibility that it provides. For instance, the number of each of multiple types of fluorophore can be stoichiometrically controlled. Sidechains can also be controllably added for attaching targeting biomolecules. Conversely, SPs are rigid and conjugated fluorescent polymers, which were originally developed for optoelectronic applications. SPs are more prominent than PPFs in the preparation of FPNs and many different colours of SPs are commercially available.93 A subclass of FPNs are polymer dots (Pdots), which are brightly fluorescent spherical NPs that range in size between 5 to 500 nm diameter.94 They are typically composed of fluorescent π-conjugated semiconducting polymers that are stabilized in an aqueous environment. Commonly used semiconducting polymers include the polyfluorenes (e.g. PFO), poly(phenylene vinylene) (e.g. Cyano-poly(phenylene vinylene) (CN-PPV)), and fluorene-based copolymers (e.g. poly[{9,9-dioctyl-2,7-divinylene-fluorenylene-alt-co-{2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylene}] (PFPV), and poly(9,9-dioctylfluorene-alt-benzothiadiazole) (PFBT)).94 Pdots exhibit characteristics that are favourable for bioanalytical applications, including high fluorescence brightness, good photostability, low cytotoxicity, nonblinking behavior, and fast emission rates.94   The Chiu Research Group defines Pdots as FPNs that are comprised of SPs with a weight or volume fraction of at least 50%, or ideally more than 80%, with particle sizes that are comparable to QDs, which are typically less than 30 nm.94 In this thesis, a Pdot is defined as a spherical 33  nanoparticle that is < 100 nm in diameter and consists of a high mass-fraction (> 50%) of fluorescent polymer (e.g. SP or PPF). Pdots are synthesized using one of two methods: mini-emulsion or nanoprecipitation. Both processes utilize similar techniques but differ based on how the particles are stabilized.95 The mini-emulsion method relies on the formation of oil-in-water droplets, whereby SPs or PPFs that are dissolved in an aprotic solvent, emulsifies in an aqueous solution consisting of amphiphilic surfactant molecules such as polysorbates (e.g. Tween). The emulsification process is accelerated by vigorous stirring and/or ultrasonication. The evaporation of the aprotic solvent results in Pdots that are stabilized with the surfactant molecules in an aqueous environment. In particular, the surfactant molecules situate their hydrophobic tails within the polymeric core, while the hydrophilic head points towards the aqueous hydrophilic environment. The mini-emulsion method typically results in Pdots that are 40–500 nm in diameter. In contrast, the nanoprecipitation method differs as the polymers are dissolved in a water-miscible solvent such as THF. The polymeric solution, consisting of an amphiphile, and fluorescent SPs or PPFs is injected into water under sonication, resulting in a decrease in the solubility of the polymers. Consequently, the formation of nanometer sized precipitates mediated by hydrophobic interactions occurs. The nanoprecipitation method usually results in Pdots with diameters in the 5–100 nm size range.64 It is believed that during the nanoprecipitation process, there is significant oxidation that occurs within the polymers, resulting in negatively charged chemical defects (e.g. carboxyl groups) which act to stabilize the particles in aqueous solutions.95 Pdots that are stabilized via electrostatic repulsion are prone to colloidal destabilization at low pH and high ionic strength. Furthermore, nonspecific binding to proteins and various surfaces (e.g. hydrophobic, cationic) is a common challenge with Pdots, sometimes even when colloidally stabilized by PEGylated amphiphiles.  The optical properties of Pdots are largely dependent on the type of SP or PPF utilized during its synthesis. When a photon is absorbed by a Pdot, an electron undergoes a  π- π* transition from the HOMO to the LUMO. The absorbance spectra of Pdots is typically quite broad (FWHM 50–200 nm),64 spanning the UV to visible wavelengths, and with multiple electronic transitions. Fluorescence occurs when an electron undergoes a π*- π transition from the LUMO to the HOMO with the release of a photon. The emissive properties of Pdots span the visible range between the blue (~450 nm) and near-IR (~720 nm) wavelengths and are typically broad (FWHM 50–100 nm).64 A main advantage of Pdots in comparison to other fluorescent 34  materials (e.g. organic dyes, fluorescent proteins, and even QDs) is their exceptional brightness. The reported absorption coefficients range between 107–1010 M–1 cm–1.64  Pdots are of growing interest a number of bioanalytical applications, including immunofluorescent labeling of cells, in vivo imaging, single-particle tracking studies, drug and gene delivery, and energy transfer-based sensing.94 Recently, Gupta et al. demonstrated that the exceptional brightness of Pdots enabled detection of lower concentrations of analyte in a prospective lateral flow assay than QDs.96   1.4.6 Single-chain polymer nanoparticles Single-chain polymer nanoparticles (SCPNs) are a diverse family of soft-matter-based materials that are prepared via the collapse of a single polymer chain mediated by intramolecular interactions and bonding. The chemical composition and functionality of the polymers provides specific properties to the SCPNs; for example, multiple pendant dyes may be conjugated to impart fluorescence properties to the final NP. The synthesized SCPNs can range in diameter between 1–20 nm.97,98 Multiple strategies are utilized to induce chain-collapse, including covalent and non-covalent strategies. Covalent methods can be further categorized into homo-functional chain collapse, hetero-functional chain-collapse, and cross-linker mediated chain-collapse. An example of homo-functional chain-collapse is the metathesis of identical olefin functional groups within the single polymer chain.99 An example of the hetero-functional chain-collapse method is the orthogonal “click” reaction that occurs between azide and alkyne pendant groups within the polymer chain, typically mediated by a copper catalyst.100 An example of cross-linker mediated chain-collapse is the cross-linking of isocyanate groups within a polymer chain via a diamine molecule.101 In contrast, non-covalent based methods are designed to mimic the folding mechanisms observed in proteins, taking advantage of π-π interactions and hydrogen bonding. An example of non-covalent SCPN formation is the use of self-complementary 2-ureido-pyrimidinone (UPy) groups that were pendantly bound to a norbornene polymer backbone and protected with UV responsive 2-nitro-benzyl groups.102 Upon irradiation with UV-light, the UPy groups spontaneously associated, forming globular NPs ~20 nm in diameter.  35  Novel fluorescent materials that are comprised of many luminescent reporters within a single particle vector are of interest due to their superior brightness in comparison to their single reporter counterparts. Greater per particle brightness can be beneficial in sensing target biomolecules (e.g. proteins, nucleic acids) or whole cells (e.g. mammalian cancer cells) that are substantially low. Moreover, brighter particles can be beneficial when combined with applications that utilize non-optimal but low-cost instruments, such as a smartphone for analyte detection in point-of-care (POC) diagnostic applications. SCPNs comprise many fluorescent dyes conjugated along a single-polymer chain. The versatility of SCPNs in terms of synthesis, construction, and properties has led to a rapid increase in the number of applications in which they have been used in, including drug delivery103 and biological imaging.104   1.4.7 Non-specific adsorption on nanoparticles In bioanalytical applications, the non-specific interactions of nanoparticles with biological molecules and interfaces (e.g. tissues, cells, proteins, nucleic acids) can result in unwanted labeling and false results. The physicochemical properties of both a nanoparticle and an interface or biological molecule or matrix determines the extent of non-specific adsorption. The types of interactions that may cause non-specific binding include hydrophobic interactions, electrostatic attractions, van der Waals forces, hydrogen bonding, coordinate bonding, and salt bridges.105,106  For example, with proteins, hydrophobic forces are often a key contributor to non-specific binding, as the hydrophobic patches of a protein may strongly adhere to hydrophobic patches at a nanoparticle interface. Cells are another possible biological interface with which nanoparticles can non-specifically interact. Cell surfaces have a net negative charge due to the phosphate groups found within the phospholipid bilayers. Furthermore, patches of cationic proteins or receptors are prevalent on the cell surface. The charge on the cell surface provides a mechanism for coulombic attraction between charged (e.g. positive or negative) particles and the cell surface.  Non-specific interactions between nanoparticles and biological substances in a sample matrix can lead to a variety of detrimental outcomes. For example, potential consequences of protein non-specific binding to nanoparticles include a reduction of colloidal stability and a reduction or loss of biological activity for nanoparticles that are functionalized with targeting biomolecules (e.g. 36  antibodies, aptamers, peptides).106 In the context of cellular immunofluorescent labeling, the ability to specifically label cell-surface antigens will be compromised if non-specific interactions with the cellular surface or nanoparticle internalization occurs, with reduced contrast between antigen-positive and antigen-negative cell types. Furthermore, these non-specific interactions can lead to a reduction in signal-to-background ratio if the nanoparticles adsorb to the substrates (e.g. glass, plastic) used for imaging cells. These outcomes ultimately contribute to a reduction in specificity and sensitivity for the detection or qualitative analysis of biological targets.   The interactions between a nanoparticle and an adsorbate can often be mitigated by coating the nanoparticle with an uncharged hydrophilic polymer (e.g. PEG, dextran) or a zwitterionic material.107,108 Such coatings can mask hydrophobic patches of a nanoparticle surface and prevent strong electrostatic interactions due to their lack of charge or net neutral charge. These coatings also demonstrate a high capacity for water solvation, which acts as a barrier for non-specific interactions with an adsorbate.109,110 Other environmental factors, such as changes in pH or temperature, can also affect the how nanoparticles interact with various biological components.  1.5 Point-of-care diagnostics  The past decade has seen a rapid increase in the development of tests that are aimed to provide quick diagnostic results in a nonlaboratory setting. The main objective of these POC diagnostic tests is to reduce the time and cost required to obtain medically useful results. Tests are performed at the real or metaphorical bedside of the patient, ideally within minutes, rather than sending the samples to be analyzed to a centralized laboratory, where wait times may vary from hours to days or even weeks. These tests must be readily available at low cost per test, portable, robust, provide reliable results with good analytical figures of merit, and, in many cases, be simple enough to be performed by non-specialized personnel.   In developed countries, improvements in health care efficiency are possible with the deployment of POC diagnostic technologies and can help address gaps in health care quality between urban and rural or remote communities. In another context, the accessibility of health care in developing countries is an ongoing challenge because of economic limitations, lack of infrastructure, and lack 37  of suitably skilled personnel, among other obstacles. These shortages make these areas ill-equipped to deal with infectious diseases. Rapid screening via POC diagnostic tests can be lifesaving in these areas and improve overall health care without requiring significant amounts of new infrastructure.   POC-like diagnostic technologies are also useful for fields beyond health care. In these contexts, the tests are commonly referred to as point-of-need (PON). PON tests have been developed for a wide variety of applications, including food and water quality assessment, agriculture and aquaculture, environmental testing, and public safety and security.  The target analytes that are detected with POC/PON tests can be grouped into three main categories: (i) proteins, (ii) nucleic acids, and (iii) small molecules.111 Arguably, cancer-related cells and pathogens are another category of target analyte; however, these targets are typically detected via a specific protein biomarker. For example, the HER2 antigen is overexpressed on the cell surface of certain breast cancers and this can be exploited for specific labeling. Common pathogens include viruses, fungi, bacteria, parasites, protozoans, and prions.  The most successful format for POC/PON tests is arguably the LFA. The LFA was introduced in 1988 by Unipath111 and is a paper-based platform that is capable of rapid detection and quantification of analytes within complex mixtures with rapid readout results. The implementation of LFAs has been achievable due to their low cost and ease of production, and these tests are typically used within hospitals, doctor’s offices, and clinical laboratories. LFAs can measure analytes in a wide variety of biological matrices including serum, plasma, whole blood, sweat, saliva, and urine.112  LFA devices consist of a series of overlapping membranes that are mounted on a backing card, which provides better stability during handling.112 At one end of the strip is an absorbent sample pad, which is embedded with buffer salts and various surfactants, which help prevent non-specific interactions between the sample and the membrane, and to maximize analyte binding with the capturing reagents on the membrane. Through capillary action, the sample migrates along the strip, towards the conjugate pad, which consists of target specific antibodies that are conjugated to 38  deeply colored labels. Commonly utilized labels include colloidal gold nanoparticles, dyed latex microspheres, and, more recently, fluorescent particles. After releasing from the conjugate pad, the targeting antibodies bind with the analytes in the sample as they migrate with the fluid flow towards the detection zone, which consists of a test line and a control line. The test line and control lines are covalently coupled with antibodies that are specific to the analyte and the IgG antibody respectively. During a test, the labeled antibody binds to the analyte. At the test line, immobilized antibody binds to another epitope on the analyte. A positive test results in the form of fluorescence (with fluorescent labels) or visible colour (with gold nanoparticles or dyed microspheres) from the sandwich complex formed between the label, analyte, and the test-line immobilized antibody. The control-line indicates that there were no problems with liquid movement or degradation of the antibodies. A test is deemed viable when a response is generated at this line upon recognition of the antibody-label complex via control-line immobilized Anti-IgG antibodies. The capillary forces that drive flow of the liquid sample are maintained by an absorbent pad that is situated at the opposite end of the sample pad. One of the most common lateral flow assays is the pregnancy test, which is an assay for human chorionic gonadotropin (hCG). The readout of an LFA can be by eye, as is the case for a pregnancy test, or, in some cases, with a specific readout instrument. Assessment of results by eye is cheaper, faster, and easier, as no additional instrumentation is required; however, quantitative results must be evaluated with a readout instrument.  1.5.1 Smartphone-based diagnostics The rapid development of smartphone technology has made this consumer device attractive as a prospective instrument for POC diagnostic tests. In this context, smartphones typically serve one of two purposes: (i) as readout device or (ii) as a user-friendly interface for a dedicated instrument.113 Smartphones are miniature computers, with advanced operating systems, internal memory, a user-friendly interface, and wireless connectivity supplementing high-quality back and front facing cameras—all bundled within a handheld device.  1.5.1.1 Smartphone components Smartphone cameras work in a similar fashion to standalone digital cameras in the sense that they utilizes lenses to capture light; a CCD or, more commonly, a CMOS image sensor to record the spatial and color distribution of light intensity; and a data processing unit to convert the acquired 39  data into an image that mimics what is seen by the human eye. Where smartphones typically differ from a digital camera is the size of their image sensors. Smartphones are designed to have smaller sensors to enable a smaller housing, typical of a slim model. The data from the image sensor comes from an array of thousands of photodiode-like detectors, each corresponding to an individual pixel as shown in Figure 1.7A. Many smartphones come with an infrared (IR)-blocking filter, which prevents unwanted IR light from distorting an image. These IR-filters are included to mimic the vision of the human eye, which lacks sensitivity in the IR wavelength range. The inclusion of an IR-filter, however, has a negative impact on the camera’s sensitivity, as it is blocking out light that may be captured by the camera’s sensor. In front of the IR-filter is a Bayer mosaic filter, which is applied over top of the photodiodes to filter the incoming light as either red, green, or blue. The Bayer mosaic is conventionally used in smartphones, as its repetitive 2×2 grid consisting of 50% green, 25% red, and 25% blue optical filters is designed to mimic what the human eye visualizes.114 The analog signals from the individual photodiodes undergo demosaicing and the underlying electronics convert the analog signal into a digital image. The spectral sensitivity of a smartphone camera sensor depends on the wavelength of light being captured. As shown in Figure 1.7B, the spectral sensitivity for blue, green, and red light peaks around ~450 nm, ~520 nm, and ~620 nm, respectively, and some wavelengths of light will be concurrently detected in multiple colour channels.  Modern smartphones also come equipped with a white light-emitting diode (LED), which typically functions as a flash for imaging in low-light conditions. White LEDs are typically made by applying a phosphor coating to blue LEDs. Upon direct excitation by the blue LED, broad spectrum visible light is emitted by the phosphor. These LEDs can be used in spectroscopic applications requiring a broadband illumination source.  The ambient light and proximity sensor on most smartphones are utilized for automatically adjusting the screen brightness in response to the amount of ambient light being captured by the sensor. The proximity sensor shoots IR light at an object and depending on the amount of light that is reflected towards the detector, can provide information about the distance of the object from the detector. This feature is useful for automatically shutting off the screen when a user is making a call with the phone held close to their face. These light sensors also be used in spectroscopic applications that require a photon detector.115,116 In addition, there are Android-based applications 40  available which allow the user to measure light intensity in lux units (e.g. Light Meter, Lux Meter, Light-O-Meter). Additional sensors on the smartphone such as the accelerometer and gyroscope provide information about its orientation in space and time.   The connectivity offered by smartphones is perhaps one of its greatest advantages. Most smartphones include medium to short-range connectivity devices such as Bluetooth and NFC, which are useful for data transmission with other nearby devices. Smartphones can also be used to transmit data globally through a Wi-Fi connection or a cellular network. The data transmission can be coupled with the Global Positioning System technology to enable geographical and temporal tagging of data measured on a smartphone.   Smartphone batteries, which can contain up to 3300 mAh of rechargeable energy, can provide a significant source of power for peripheral components such as LEDs or motors. Through the microUSB port, up to 10 W of power can be drawn from the battery, at a potential of 5 V and up to 2.0 amperes. A typical LED operates at a voltage between 1.8 to 3.3 V, with up to 20 milliamperes of current, which is well-below the maximal output that can be drawn through the USB port.   41   Figure 1.7 (A) Simplified schematic of a CMOS image sensor. (B) Spectral sensitivity of a typical CMOS image sensor without (black) and with RGB colour filters (coloured lines). The typical blocking region of an IR filter is also shown. Figure adapted with permission from ref.111  1.5.1.2 Image processing The smartphone camera is the most commonly utilized component within the smartphone for bioanalytical detection. In scenarios where analyte quantification is important, image pixel intensity can be analyzed. Most modern smartphones acquire images with a 24-bit colour-depth, which is a total of the sum of each 8-bit colour channel (e.g. red, green, blue). Within each colour channel, the pixel intensity is recorded as one of 256 brightness levels (e.g. 0–255), allowing for an RGB image that can display up to 16 777 216 different colours. The images typically obtained with a smartphone camera are in JPEG format, which is a compressed image format that has different parameters automatically adjusted, such as brightness and white balance. These auto-adjusted features alter the pixel data, as the smartphone attempts to adjust the image for a desired effect (i.e. the observed pixel intensity is not what was originally measured). However, many 42  modern smartphones are capable of capturing images in an unprocessed, or raw format. This format is advantageous as it provides the measured pixel intensity, without distortion from the smartphone. There are applications, both built-in or available (e.g. Apple AppStore or GooglePlay), that provide precise control over imaging parameters such as ISO, shutter speed, and white balance, and offer capability for imaging in raw format. In most cases, the raw pixel intensity is more directly proportional to intensity of light detected. Raw images can be analyzed in-situ, with custom built smartphone applications, or remotely using image processing software such as ImageJ, MATLAB, or Adobe Photoshop. Target analytes are typically quantified based on a calibration curve that relates pixel intensity to analyte concentration for the specific smartphone imaging set-up and assay format.  1.5.1.3 Utility of smartphones in POC diagnostics POC diagnostic tests that utilize smartphones as a detector or user interface are abundant and diverse in the scientific literature. Smartphones have proven useful for the detection of a variety of small molecule analytes including hormones,117 electrolytes,118 vitamins,119 and amino acids. In the context of public safety, smartphone-based assays have been developed for small molecule analytes such as explosives,120 toxins,121 and illicit drugs such as cocaine and methamphetamine.122 Similarly, the detection of nucleic acids is important for the diagnosis of diseases, and for the identification of certain pathogens and genetic predispositions.111 Highly sensitive detection of disease-specific nucleic acids on smartphone-based devices has been reported including those specific for Human Papillomavirus (HPV) and dengue virus type I.123 The detection of protein biomarkers using a smartphone-based detector include analytes such as antibodies,123 hormones, enzymes,111,124,125 cytokines, and viral nonstructural proteins126. Recently, the detection of exosomes has garnered considerable interest as a next-generation biomarker for noninvasive early cancer diagnosis,127,128 and a LFA with smartphone-based detection has been developed for this purpose.129 The smartphone has also been leveraged in applications that analyze larger whole-cell organisms such as bacterium and mammalian cancer cells.130 A summary Table 1.2 below provides more details about the cited literature. 43  Table 1.2 Representative examples of smartphone-based point-of-care diagnostics. Target(s) Class (e.g. small molecule, nucleic acid, protein, vesicle, whole cell) Readout mode Amplification (Y/N) Substrate Label Type Figures of Merit Cocaine, methamphetamines ref.122 Small molecule Colorimetric N Centrifugal microfluidic platform Simon’s Reagent Detection limits: 0.25a, 0.75b  mg mL-1 Aflatoxin B1 ref.121 Small molecule Fluorescence N Molecularly imprinted polymer membrane Intrinsic polymer fluorescence Detection limit: 20 ng mL-1 HIV1-p17, dengue virus type I specific antibodies ref.123 Proteins Biochemiluminescent N In solution: in microwell plate Luciferase, mNeonGreen Detection limit: 10 pM Enzymes ref.124 Proteins Fluorescence N Paper in PDMS QDs Detection limit: 18 NIH units ml-1 Human Papillomavirus (HPV) DNA ref.131 Nucleic acid Chemiluminescent Y Microfluidic cartridge Horseradish peroxidase Sensitivity: <10 aM Exosomes ref.129 Vesicle Colorimetric N Lateral Flow Strip Au@Pd nanopopcorn Detection limit: 1.4 × 104 exosomes µL-1 Giardia lamblia cysts ref.132 Whole cell Fluorescence N Filter membrane Antibody-fluorescein conjugate Detection limit: 12 cysts per 10 mL Staphylococcus aureus, Klebsiella pneumoniae, Enterobacter, Citrobacter, and a panel of breast cancer cells (BT474, MCF7, HC1937) ref.130 Whole cell Upconversion luminescence N In solution: drop-cast on plastic slide UCNPs Differentiated bacterial samples, identified patients at high-risk for malignancy Notes: a Cocaine, b methamphetamines, Y = Yes, N = No 44  1.6 Overview and contributions of this thesis   This thesis evaluates novel and emerging luminescent nanoparticle materials—including SCPNs, Pdots, QDs, and composite materials such as silica nanoparticle-quantum dot supra-nanoparticles (SiO2@QDs) and magnetic nanoparticle-quantum dot supra-particles (MNP@QDs)—primarily for applications in cellular immunolabeling and imaging. The materials that are investigated within this thesis are illustrated in Figure 1.8, alongside the utilized immunoconjugation strategies in Figure 1.9. In addition to the evaluation of all of these materials for imaging by fluorescence microscopy, the high per-particle brightness of some of these materials (e.g. Pdots, SiO2@QDs, MNP@QDs) is evaluated for smartphone-based imaging of labeled cancer cells in proof-of-concept diagnostics. TACs are also evaluated as a novel and advantageous immunoconjugation strategy for some of the luminescent nanoparticles. To date, TACs have primarily been utilized in cellular isolation applications with magnetic nanoparticles, but not with fluorescent nanoparticles.     Figure 1.8 Luminescent nanoparticles evaluated for immunolabeling: Single Chain Polymer Nanoparticles (SCPN), HMAT-ODA Pdot, F8BT Pdot, CNMEHPPV Pdot, SiO2@QD, and MNP@QD. Approximate hydrodynamic diameters: SCPNs 26 nm, Pdots 50–70 nm, QDs 2–10 nm, SiO2@QDs ~110 nm, MNP@QDs ~250 nm. The inset table shows the approximate brightness values for the luminescent nanoparticles. Brightness values are approximations based on the calculated product of quantum yield and molar absorption coefficient for the respective materials, with excitation at 405 nm for all materials except SCPNs (450 nm). The ranges for the brightness of QDs and Polymer dots are a function of nanoparticle size and material. 45    Figure 1.9 Strategies for immunolabeling fixed SK-BR3 cells with fluorescent nanoparticles.  This thesis is divided into nine chapters, including this introductory chapter, plus appendices, which contain additional experimental methods and results. With one exception, a common theme across the chapters is the development and evaluation of luminescent nanoparticles and immunoconjugates for extracellular labeling.  Chapter 2 describes the development of single QDs coated with dextran (Dex-QDs) for bioanalytical applications. Although widely used for the stabilization of iron-oxide magnetic nanoparticles, dextran is a scarcely utilized coating for fluorescent nanomaterials. This chapter addresses the design of multiple dextran ligands for coating QDs. For the resulting dextran-coated QDs, colloidal stability is evaluated; biocompatibility is assessed in the form of non-specific binding to proteins and live cells, and cellular viability assays; and proof-of-concept applications in bioanalysis are demonstrated, including novel TAC-mediated cellular immunolabeling with QDs.   Chapter 3 describes the development and photophysical characterization of fluorescent SCPNs for specific extracellular labeling of breast cancer cells. It is the first example of specific 46  extracellular labeling with SCPNs. Prior examples of cellular labeling with SCPNs were non-targeted and limited to intracellular staining.   Chapter 4 describes the development of novel blue-emitting Pdots based on a non-π-conjugated polymer pendantly modified with HMAT-ODA dyes. Pdots have been conventionally synthesized using fluorescent semiconducting π-conjugated polymers. Novel emerging fluorescent materials comprised of pendant dyes is advantageous as it allows for greater synthetic flexibility, with the capability of tuning the polymer color with different combinations of dyes. HMAT-based emitters have displayed properties that are favourable for bioanalysis and bioimaging, including good resistance to photobleaching, high quantum yield, high brightness, and the capability for two-photon excitation.   Chapter 5 describes the development of dextran-coated Pdots (e.g. F8BT/Dex, CNMEHPPV/Dex) that are composed of π-conjugated polymers. The dextran-coated Pdots are utilized for extracellular labeling of breast cancer cells mediated by TACs. Analogous to the QDs in Chapter 2, TACs have not been previously utilized with Pdots.  Chapter 6 describes the development of silica nanoparticle-quantum dot supra-nanoparticles coated with dextran. These supra nanoparticles were utilized for TAC-mediated extracellular labeling of breast cancer cells, with both smartphone-based and microscope-based fluorescent imaging. The high per-particle brightness of these supra-nanoparticles increased the capability of smartphone-based imaging for proof-of-concept diagnostics.  Chapter 7 describes the development of composite magnetic nanoparticle-quantum dot supra-nanoparticles for specific breast cancer isolation and detection on a 3-D printed smartphone-based imaging platform. A modular design for the preparation of multiple colours of fully self-assembled composite magnetic-luminescent particles (MNP@QDs) is demonstrated. In conjunction with smartphone-based imaging, these particles provide a rapid method for the specific detection of cancer cells in proof-of-principle diagnostic tests.   47  Chapter 8 describes the development of ligand-functionalized paper substrates for immobilization of QDs and metal nanoparticles via coordinate bonds, including patterning of QDs by microcontact printing. Paper-based analytical technologies are attractive due to the low cost, abundance, and amenability of the paper toward chemical functionalization and other processing. The research in this chapter is the foundation upon which hybrid paper-and-nanoparticle-based technologies can be developed, including paper-based molecular diagnostics.   Chapter 9 summarizes the previous chapters and discusses possibilities for future research that builds from the results in this thesis.  Overall, this thesis presents novel and important contributions to the broad field of bioanalysis and imaging with luminescent nanoparticles, and particularly cellular immunolabeling, imaging, and detection using smartphone technology. It is anticipated that the research in this thesis will ultimately lead to the development of point-of-care molecular and cellular diagnostics supported by smartphone-enabled devices.   48   Dextran-functionalized semiconductor quantum dot-bioconjugates for bioanalysis and imaging   This chapter is an adaptation of Rees, K.; Tran, M.V.; Massey, M.; Kim, H.; Krause, K.D.; Algar, W.R., Dextran-Functionalized Semiconductor Quantum Dot Bioconjugates for Bioanalysis and Imaging. Bioconjugate Chem. 2020, 31, 861–874, with permission from the American Chemical Society (Copyright 2020). Refer to the Preface for full details of author contributions. Unless indicated in figure captions, I had a sole or significant role in obtaining the data presented in this chapter.   2.1 Introduction  Semiconductor QDs have many properties that are desirable in bioanalytical applications. By nature of their inorganic structure, they display greater resistance to photobleaching than other commonly utilized fluorescent labels, including organic dyes. Their PL emission spectra are narrow with a FWHM of 25–30 nm, and their absorption spectrum are broad (FWHM > 150 nm).64 These absorbance and emission properties make them ideal for applications where multiplexing is desired because multiple colors of QDs may be excited simultaneously without significant optical crosstalk.   The surface of QDs can also be used as a scaffold, including functionalization with many stabilizing ligands and conjugation with biomolecules for bioanalytical applications. As-synthesized QDs are typically stabilized with hydrophobic surfactants in organic solvents, making them unsuitable for bioanalytical applications. QDs are rendered hydrophilic via one of two methods: amphiphilic polymer overcoating, or ligand exchange with hydrophilic molecules. The amphiphilic polymer overcoating method is advantageous as it results in colloidally stabilized QDs that have many functional groups (e.g. carboxyl, amine) amenable to conjugation with biomolecules such as antibodies, nucleic acids, and peptides. Furthermore, QDs functionalized via this route are typically brighter and more robust. A disadvantage of this overcoating approach is that it typically results in QDs that have a large hydrodynamic radius, making them unsuitable for 49  certain cellular labeling applications where size may limit access to biological environments or structures.26 In contrast, the ligand exchange method relies on the displacement of surface hydrophobic ligands with hydrophilic ligands through a mass-action process. Many of the first examples of ligand-exchanged QDs used thioalkyl acids such as dihydrolipoic acid (DHLA),133 mercaptoacetic acid (MAA),134 and mercaptopropionic acid (MPA).135 These ligands colloidally stabilize the QDs through electrostatic repulsion between the carboxyl groups at the distal ends of the ligands. A significant disadvantage of these ligands is their propensity to bind non-specifically to proteins in common sample environments such as serum, and cell suspensions. Non-specific protein binding can lead to false positive or false negative results, unwanted cellular labeling, and inactivation or inhibition of the desired biological activity.  The development of novel QD ligands that provide good colloidal stability and display anti-fouling properties is an active research area.136 The strategies that are implemented when designing these ligands include properties such as steric bulk, neutral charge, and strong hydration.137 Zwitterionic ligands are small, polar, molecules that display a net neutral charge in solution. The net neutral property of these ligands prevents electrostatic interactions with proteins in solution. The zwitterionic nature of these ligands also affects solvation, typically resulting in a more densely hydrated nanoparticle, which acts as a barrier for protein adsorption. Other ligands that exhibit many anti-fouling characteristics are poly(ethylene)glycol (PEG)-based ligands. These PEG ligands have typically been terminated with DHLA as a method to anchor to the inorganic surface of the QDs.138 Many derivatives of PEG have been developed, and may incorporate functional groups at their terminus that are suitable for bioconjugation (e.g. amine, carboxyl). A disadvantage of the PEG-based ligands is that the chemistry is limited to the terminal ends of the polymer, rather than the backbone. Despite its limitations, the PEG ligand remains widely used for many materials in addition to QDs.139,140  Dextran presents many properties that are desirable in a biocompatible nanoparticle coating, including anti-fouling, hydrophilicity, biodegradability with slow digestion rates by human enzymes, low immunogenicity and antigenicity, chemical stability at physiological pH, low cost, and commercial availability at a variety of molecular weights.141–144 Dextran materials have been used extensively with iron oxide nanoparticles as a stabilizing coating in cellular isolation and 50  magnetic resonance imaging studies.75 Furthermore, commercial sources of dye-conjugated fluorescent dextran are used as intracellular tracers due to their biocompatibility145 and anti-fouling properties. An advantage of dextran is that it is not chemically inert along its backbone, lending it synthetic flexibility when developing ligands for QDs. With iron-oxide nanoparticles, dextran is typically bound to the surface of the particles primarily through coordinate interactions that are formed between the surface iron atoms and hydroxyl moieties on dextran. 75 QDs utilize ligands with different functional groups such as imidazole, thiols, and amines.136 These ligands form dative coordinate bonds with the ZnS shell of the QD.  In this chapter, several strategies for the development of dextran-functionalized QDs (Dex-QDs) are reported. Several dextran ligands, utilizing both DHLA and 1-(3-aminopropyl)imidazole (API) anchoring groups were prepared. QDs coated with dextran ligands were compared to QDs coated with small-molecule ligands (e.g. DHLA, GSH) in studies investigating colloidal stability, protein and cellular anti-fouling behaviour, and cellular viability. Bioconjugation of QDs via covalent chemistry, or self-assembly through TACs was demonstrated. Furthermore, proof-of-concept bioanalytical applications of Dex-QDs were tested, including use as a ratiometric pH sensor, immunofluorescent breast cancer cell labeling, and other proof-of-concept applications. Overall, the Dex-QDs demonstrated excellent anti-fouling properties and wide applicability in a variety of bioanalytical applications.  2.2 Results and discussion  2.2.1 Synthetic strategies  Figure 2.1A shows the structures of the four dextran ligands in this study: dextran modified with a terminal dithiol group (D-t-DHLA), dextran modified with multiple pendant dithiol groups via one of two approaches (D-p-DHLA and D-p-DHLAm), and dextran modified with multiple pendant imidazole groups (D-p-API). The ligands were usually synthesized on a 1–2 g scale and differed in the type of anchoring group that bound to the inorganic surface of the QDs (dithiol or imidazole), the number and position of these modifications (multiple pendant or single terminal), and, in some cases, the molecular weight (MW) of the dextran (~6 or ~10 kDa, estimated MW range of ± 1.5 kDa or less). The DHLA and API parts of the abbreviations are derived from lipoic 51  acid (dihydrolipoic acid if in its reduced form) and 1-(3-aminopropyl)imidazole, respectively, which were starting materials for the dextran modifications. Intermediates in the synthesis of the final dextran ligands are denoted with the suffixes -CHO and -NH2, representing the partially oxidized and amine-functionalized dextrans, respectively. The abbreviations D6 and D10 are used to indicate dextran molecular weights of ~6 and ~10 kDa, respectively.  The rationale for exploring various ligand designs was multifaceted. A terminal anchoring group was expected to produce a dextran coating that was thicker and more densely packed than possible with pendant anchoring groups; however, multiple pendant anchoring groups were expected to more stably attach to the QD through multidentate coordination. Dithiol groups bind more tightly to QDs than imidazole groups but are prone to re-oxidation and often reduce quantum yields. 146–148 Higher molecular weight dextran was expected to provide better colloidal stability at the cost of a larger hydrodynamic size. Various ligand designs were thus evaluated in anticipation of trade-offs between advantages and disadvantages.       52    Figure 2.1 (A) General structures of the dextran ligands. The R group may be another modification, an unreacted aldehyde (or the corresponding hydrate), or have cyclized with the secondary amine of the shown modification. (B) Cartoon schematics (left) and scale illustrations (right) of the two general QD-functionalization strategies: terminal (reducing end) modification of dextran with an anchoring group, and pendant modification of dextran with multiple anchoring groups. For the illustrations, the translucent sphere around the QD approximately corresponds to the measured hydrodynamic radius, rH. Multiple possible arrangements for the pendant-functionalized dextran are shown. (C) Cartoon schematic of a TAC (left) and scale illustration (right) for a TAC bound to a dextran-functionalized QD. Scale illustrations contributed by Katherine Krause.     53  2.2.2 Ligand characterization  Unambiguous characterization of the dextran ligands was challenging because of the intentionally low degrees of modification, whether terminal or pendant, which aimed to preserve the native properties of dextran. Nevertheless, we characterized our ligands via 1H NMR and FTIR spectroscopy (Figure A-1–Figure A-5), and through colorimetric tests. The colorimetric tests were only qualitative or semi-quantitative but had the advantage of being insensitive to native dextran. Both ninhydrin and 2,4,6-trinitrobenzenesulfonic acid (TNBS) tests for primary amines gave the expected results: negative results for dextran and its derivatives prior to modification with hexamethylenediamine (HMDA), positive results after modification with HMDA, and approximately negative results after HMDA was reacted with lipoic acid (LA)-succinimidyl ester (see Figure A-6 and Table A-1). Likewise, Ellman’s test (with 5,5’-dithio-bis-[2-nitrobenzoic acid], DTNB) for thiol groups gave the expected positive result for dextran modified with DHLA and negative results with HMDA and LA modifications, and with both native and oxidized dextran (see Table A-1). It should be noted that modification of dextran with 100% yield was not necessary because the ligand exchange process for functionalizing the QDs preferentially selected for fully modified ligands (vide infra).  2.2.3 Characterization of dextran-functionalized QDs As-synthesized, hydrophobic QDs in non-polar solvent were functionalized with dextran ligands in two steps because the dextran was, for preparative purposes, only soluble in aqueous buffers and therefore unable to access the hydrophobic QDs. The first step was ligand exchange and aqueous phase transfer with histidine (His) to form His-functionalized QDs (His-QDs). The second step was aqueous exchange of the His ligands with a dextran ligand. Dex-QDs are denoted with the abbreviations for the ligands; for example, (D6-t-DHLA)-QD600 is a QD with peak PL emission at 600 nm and coated with 6 kDa dextran modified at its terminus (i.e. reducing end) with a dithiol group. Figure 2.1B shows cartoon depictions and simple scale illustrations of Dex-QDs. These illustrations are intended to help in conceptualizing the relative sizes of the QDs, the dextrans, and their bioconjugates. They are not rigorous molecular models. Notably, radially extended dextran does not correspond to the measured hydrodynamic diameters of the Dex-QDs (vide infra), and the distribution of anchoring sites for pendant-modified dextran may lead to a variety of conformations for the QD-bound dextran chains. 54  2.2.3.1 Surface chemistry Infrared absorption spectra for the various Dex-QDs (Figure A-7) showed the characteristic C–O stretch at 1010 cm–1 for dextran and no resonances that were indicative of residual His ligands. Colorimetric anthrone assays on the functionalized QDs gave positive results for dextran as shown in Figure 2.2.149 Given the unrealistically small footprint that dextran would need on the QD to reach the measured number of ligands per QD, we suggest that the relative number of dextran ligands per single QD are overestimated by the assay. There is a possibility that unbound dextran entangled with the QD-bound dextran coating and contributed to this high ratio.  Figure 2.2 Anthrone assay for determining the number of dextran ligands per QD. The values are calculated based on quadruplicate measurements and are expressed as the mean (± 1 standard deviation).   To further confirm that the QDs were functionalized with dextran, tests were done by mixing with concanavalin A (ConA), a lectin that has a tetrameric structure. In the presence of Mn2+ and Ca2+ (aq), ConA binds to carbohydrates with high affinity.150,151 As shown in Figure 2.3, all of the Dex-QDs aggregated with the addition of ConA. Other proteins did not induce aggregation  (see Figure 2.4). 55    Figure 2.3 Confirmation of dextran-functionalization of QDs by selective aggregation with ConA. Top: Photographs of X-QD solutions without (–) ConA, with (+) ConA, and with both ConA and glucose (Glc). The white arrows indicate aggregates. DHLA-QDs were used as a non-dextran control and Glc was used for competitive binding with ConA. Bottom: Agarose gel electrophoresis of analogous samples. The Dex-QD samples with ConA (and without Glc) aggregated and did not migrate from the wells. The dashed white line represents the well positions. The solid white line represents a gap of dark space digitally edited from the gel. The brightness of the DHLA samples (orange dashed box) was digitally enhanced for better visibility in the figure. Data was contributed by Kelly Rees.  With excess glucose present, little or no aggregation was observed with the addition of ConA, which confirmed specific binding between the ConA and dextran present on the QD surface. Control samples of DHLA-QDs aggregated in all cases after centrifugation from the effects of non-specific binding of ConA and its divalent cation cofactors. (The latter was reversible with EDTA, as in the running buffer for an agarose gel.)    56   Figure 2.4 (D10-t-DHLA)-QD600 incubated with 100 molar equivalents of dextran-specific (ConA) and non-specific (BSA, Lyz) proteins. Wells are labelled with a white dashed line. Agarose gels were prepared as a 0.5% (w/v) solution in TBE buffer. Gels were run at ~6.7 V cm–1 for 30 min.   2.2.3.2 Optical characterization The absorbance and PL excitation and emission spectra for QDs (Figure A-8) did not change significantly upon functionalization with the dextran ligands. There was a ≤ 5 nm shift in wavelength for the PL emission maximum and no change in FWHM relative to the as-synthesized hydrophobic QDs. Measured quantum yields for the various Dex-QD600 were between 9–29%, consistent with expectations from our experience functionalizing QDs with other thiol-based ligands. Considering batch-to-batch variation, Dex-QDs gave similar or slightly higher quantum yields than control samples of DHLA-QDs. Consistent with expectations, (D-p-API)-QDs generally had higher quantum yields than (D-t-DHLA)-QDs, (D-p-DHLA)-QDs, and (D-p-DHLAm)-QDs.   2.2.3.3 Size characterization QD600 had a hard size of 9.8 ± 1.3 nm, as measured by transmission electron microscopy (TEM; Figure A-9). Dynamic light scattering (DLS) and nanoparticle tracking analysis (NTA) were used to estimate the solvodynamic size of various X-QD600 samples, where X = DHLA and X = glutathione (GSH) were reference materials. Table 2.1 summarizes the number-weighted DLS results. As expected, the Dex-QDs had larger hydrodynamic sizes than QDs functionalized with the small-molecule ligands GSH and DHLA, and the (D10-t-DHLA)-QDs had a larger size than (D6-t-DHLA)-QDs. Interestingly, the pendantly-modified Dex-QDs were similar in size to the 57  terminally-modified Dex-QDs. Table 2.1 also includes data for (D6/10-t-DHLA)-QDs that were prepared with a 95:5 ratio of D6-t-DHLA and D10-t-DHLA.   Table 2.1 Solvodynamic sizes of X-QD600. Data contributed by Kelly Rees.  Diameter (nm) Ligand X DLS a NTA b D6-t-DHLA 22 ± 2 13 ± 1 D6/10-t-DHLA 22 ± 4 15 ± 1 D10-t-DHLA 41 ± 9 21 ± 1 D6-p-DHLA 26 ± 3  16 ± 1 D6-p-DHLAm 22 ± 1 14 ± 1 D6-p-API 22 ± 1  10 ± 1 DHLA 13 ± 5 -- GSH 10 ± 1 -- Notes: a Measured in carbonate buffer. b Measured in 40% v/v glycerol (aq).   For NTA with QDs, it was necessary to use buffer solutions with 40% v/v glycerol. The greatly increased viscosity (η ~0.0040 versus 0.00089 N s m–2 for water) sufficiently slowed the diffusion rate of the QDs for tracking by PL emission. GSH-QDs and DHLA-QDs were not measured by NTA as there were indications of aggregation in the 40% glycerol. The solvodynamic diameters were smaller (by about half) than those determined by DLS and were closer to the hard size of the QDs. Consistent with the DLS data and expectations, (D10-t-DHLA)-QDs had the largest solvodynamic size. The NTA results indicated that the samples were monodisperse without a secondary population of aggregates. Beyond the foregoing, we do not rely heavily on the NTA-derived sizes for two reasons: first, the DLS-derived sizes are truly hydrodynamic whereas the NTA-derived sizes are solvodynamic because of the high percentage of glycerol; second, tracking individual QDs with commercial NTA instruments is challenging (this work being one of the first examples) and further method optimization is warranted.    58  2.2.3.4 Electrophoretic mobility Agarose gel electrophoresis and capillary electrophoresis (CE) were useful for evaluating the functionalization of the QDs with dextran. Agarose gel electrophoresis has proven particularly useful for characterization of QD surface chemistry.152 Shifts in electrophoretic mobility reflect changes in relative size and charge. Moreover, a narrow band on a gel may indicate a monodisperse or relatively homogeneous sample, or surface chemistry that is static. A band that streaks may indicate polydispersity or heterogeneity, or surface chemistry that is dynamic (e.g. labile ligands). A punctate pattern in or near the sample well often indicates aggregate formation. CE is complementary to gel electrophoresis because neutral species, like Dex-QDs, will have non-zero mobility. In contrast, a neutral species will have minimal mobility on an agarose gel, which is not unambiguously different from a charged material with large size.  The agarose gel electrophoresis results in Figure 2.5A show that His-QD600 (the precursor to Dex-QD600) were negatively charged and migrated toward the anode, sometimes as a well-defined band and sometimes with streaking. The latter was symptomatic of the relatively weak binding of His to the QD. Next, precursors to the fully-modified dextran ligands were tested. The ligand exchange procedure for each precursor was analogous to that for the full ligands, save for omission of the disulfide reduction step (where applicable). QDs prepared with unmodified dextran and D-t-NH2 showed punctate patterns in the wells, indicative of aggregation, as well as significant streaking. Thus, neither the dextran nor the D-t-NH2 had sufficient binding affinity with the QD to impart colloidal stability, and rather destabilized the His binding. The addition of D6-p-NH2 did not aggregate the QDs but showed streaking from the well with less migration than His-QDs. The increased binding of D6-p-NH2 with QDs versus D6-t-NH2 likely arose from the multiple pendant amine groups of D6-p-NH2. However, as indicated by the migration towards the anode, the QDs were still negatively charged, suggesting that the His ligands were not efficiently displaced. This result suggests that the D6-p-NH2 electrostatically bound to the anionic His-QDs, as the aminated dextran would likely be partially protonated (i.e. cationic) under the gel conditions (pH 8.3), whereas the His ligands would be deprotonated. His-QDs treated with oxidized dextran showed no aggregation and fully migrated from the well, but over a lesser distance than His-QDs alone. We speculate that this result was from reaction of some of the amine groups of His with the 59  oxidized dextran, thereby retaining the negative charge but increasing the effective hydrodynamic size of the QDs.   Figure 2.5B shows that the agarose gel electrophoresis results were very different with the fully modified dextran ligands. Dex-QDs were expected to be neutral in charge and, consistent with expectations, either largely remained within the well (without signs of aggregation) or migrated slightly toward the cathode. Greater cathodic mobility was observed with D10-t-DHLA versus  D6-t-DHLA, and the mobility of QDs coated with a mixture of 95/5 mixture of D6/D10-t-DHLA was intermediate. Electroosmosis within the gel was a possible cause of the slight cathodic mobility, which has also been seen with neutral PEG-functionalized QDs.153 We also saw similar results with QD materials other than QD600 (data not shown). Overall, the agarose gel electrophoresis results confirmed that modification of dextran with terminal or pendant dithiol or imidazole groups was necessary for efficient functionalization of the QDs.   The Dex-QDs did not have sufficient mobility for reliable zeta-potential measurements, so additional evidence of their approximate charge neutrality was sought via CE. Fluorescein, an anionic dye, was included in all runs as an internal standard. Figure 2.5C shows that the elution peak of QDs functionalized with the various fully-modified dextran ligands approximately aligned with the elution peak of Rhodamine B, which was used as a neutral reference dye marker.154 This result confirmed the expected neutral charge of the Dex-QDs. 60    Figure 2.5 (A) Agarose gel of His-QDs and QD samples prepared using precursors to the fully-modified dextran ligands (i.e. no DHLA or API groups). (B) Agarose gel of His-QDs, Dex-QDs prepared with the fully-modified dextran ligands, and DHLA-QDs as a reference. (C) Electropherograms of the different QD samples with fluorescein as an anionic internal standard (eluted at ~4 min). Rhodamine B was used a neutral reference dye marker (eluted at ~2 min). The fluorescein internal standard is indicated with a black arrow. The elution times for the QDs and Rhodamine B are indicated with the coloured, horizontal arrows. Data collected in collaboration with Kelly Rees 61  2.2.3.5 Colloidal stability A desirable property in a biocompatible QD coating is colloidal stability across a wide range of pH and at high ionic strength (Figure 2.6). The colloidal stability of various Dex-QDs was compared to DHLA-coated QDs as a control. In comparison to the DHLA-QDs, the Dex-QDs demonstrated greater colloidal stability across a wider pH range (3 to 8) over a period of at least 8 weeks. In contrast, the DHLA-QDs aggregated at low pH (< 4) and, after 8 weeks, the samples that were prepared at pH < 7 had aggregated. Furthermore, high ionic strength resulted in immediate aggregation of DHLA-QDs, whereas the Dex-QDs remained colloidally stable for weeks.   Figure 2.6 (A) Photos of samples of (D10-t-DHLA)-QD600 and DHLA-QD600 in buffers at various pH after initial preparation and after 8 weeks. The photos show the QD PL under UV illumination. The white arrows indicate the presence of visible aggregation. (B) Summary plot of the approximate period of colloidal stability at pH 3–8 and at 1 M ionic strength. Data contributed by Kelly Rees. 62  2.2.4 Assessing non-specific binding  For biological applications of QDs, an important consideration is the non-specific binding with biomolecules. For example, the non-specific binding of proteins, sometimes referred to as the formation of a protein corona, can occlude the surface of QD bioconjugates to suppress or alter biological activity, and cause aggregation.155 We therefore investigated non-specific binding with Dex-QDs.  2.2.4.1 Protein binding via gel electrophoresis Non-specific binding of proteins with Dex-QDs (D10-t-DHLA, D6-p-DHLA, and D6-p-API) and DHLA-QDs (as a reference) was assessed using agarose gel electrophoresis. The proteins investigated were bovine serum albumin (BSA), bovine plasma (~3 % albumins, ~4% globulins, <1% fibrinogen),156,157 lysozyme (14.3 kDa, pI 10.7), and skim milk powder (predominantly casein, which has 4 subunits ranging between 18–25 kDa and pI 4.1–5.8). Protein adsorption was indicated by a change in the electrophoretic mobility of the QDs. As shown in Figure 2.7, the DHLA-QDs exhibited greater changes in electrophoretic mobility in comparison to the Dex-QDs, and aggregated upon incubation with lysozyme, as a result of electrostatic attraction between the DHLA and proteins. The D10-t-DHLA-QDs demonstrated the least change in electrophoretic mobility upon exposure to proteins, indicating that protein adsorption was minimized for this coating. QDs coated with D6-p-DHLA, and D6-p-API showed significantly larger shifts in electrophoretic mobility in comparison to the D10-t-DHLA-coated QDs; however, the changes were less than what was observed for the DHLA-QDs, and there was no visible aggregation. Further interpretation can be found in Table A-2 and Table A-3 within Appendix A.7. A possible limitation of this study is the possibility that weakly adsorbed proteins would get displaced from the QDs upon the application of the electric field. This would result in a similar electrophoretic mobility in comparison to unexposed QDs. Therefore, cellular nonspecific binding studies (vide infra) were conducted to complement the data shown here.  63    Figure 2.7 Agarose gels of X-QDs incubated with 1.0, 5.0, and 9.5 mg/mL protein solutions (in carbonate buffer, excluding plasma, which is percent v/v), where X = D10-t-DHLA, D6-p-DHLA, D6-p-API, and DHLA. Strong non-specific protein binding was observed with DHLA-QDs. Depending on the protein(s), some Dex-QDs exhibited little or no non-specific binding, whereas others exhibited some non-specific binding. Refer to text for details. Data contributed by Melissa Massey. 64  2.2.4.2 Binding to cells To test non-specific binding to cultured live cells, Dex-QDs were incubated with two different mammalian cell lines, A549 (human lung cancer) and MDA-MB-231 (human breast cancer). DHLA-QDs aggregated in the buffer used with the cells, so GSH-QDs were used as a reference material instead. (GSH-QDs generally exhibit lower non-specific binding and greater colloidal stability than DHLA-QDs.) Figure 2.8A-B and Figure 2.9 show that, for both cell lines, GSH-QDs exhibited the largest amount of non-specific binding, with a ~2.8-fold increase in fluorescence intensity versus the cell autofluorescence (232 ± 0.7 a.u.). Dex-QDs showed much lower non-specific binding, with only 1.1- to 1.3-fold increases in fluorescence intensity relative to the control. Across the dextran ligands, (D6-p-API)-QDs showed the most non-specific binding and (D10-t-DHLA)-QDs showed the least, consistent with the agarose gel electrophoresis results on protein adsorption and the suspected weak affinity of the coating for the QD (vide supra). Figure 2.8C shows that the superior resistance of (D10-t-DHLA)-QDs and (D6-p-DHLA)-QDs to non-specific cell binding versus GSH-QDs was largely maintained over an order of magnitude range of QD concentration. Depending on the dextran ligand, cell line, and QD concentration, the contrast in the non-specific PL signal was between 1–3 orders of magnitude.   65    Figure 2.8 (A) Representative images of live A549 cells that had been incubated with X-QD600 (50 nM), where X = GSH or D10-t-DHLA. Negative controls were cells that were not incubated with QDs. Scale bars are 50 µm. A pixel intensity calibration bar is provided. All images were acquired under the same microscope, camera, and image processing settings. Additional images are shown in Figure 2.9. (B) Average post-wash PL intensity of live A549 and MDA-MB-231 cells incubated with various X-QD samples at 50 nM for 30 min. The average cell autofluorescence (autoFL) was 232 ± 0.7 a.u. and is indicated by the black dashed line and diagonally crosshatched bars. (C) Quantitative assay of non-specific binding of X-QD to live A549 and MDA-MB-231 cells incubated with varying concentrations of X-QD600, where X = GSH, D6-p-DHLA, and D10-t-DHLA. Experiments and data analysis performed with Kelly Rees.  66   Figure 2.9 Comparison of non-specific binding between live cells and QD600 (50 nM) with various ligand coatings. Images were acquired under brightfield and fluorescence modes. (A) A549 cells. (B) MDA-MB-231 cells. Negative controls are cells that were not incubated with QDs. Scale bar = 50 µm. Exposure time = 150 ms. Images were acquired at the same microscope and camera settings. A pixel intensity calibration bar is provided on the right. Experiments and data analysis performed with Kelly Rees.  2.2.5 Microinjection and cell viability  Cellular microinjection studies were done as a preliminary test toward the feasibility of utilizing Dex-QDs as intracellular probes or sensors. In cellular microinjection studies with Dex-QDs and live A549 lung carcinoma cells, we observed no adverse morphological cellular effects (e.g. blebbing, cell shrinking) upon the microinjection of the Dex-QDs (Figure 2.10A). Furthermore, the Dex-QDs were evenly distributed within the cytosol of the cells, suggesting that these QDs remained colloidally stable. In contrast, when DHLA-QDs were microinjected into the cells, they immediately aggregated, and were not dispersed throughout the cell (i.e. remained at the injection site, see Figure 2.10B).  67    Figure 2.10 (A) Brightfield, PL, and merged images of X-QD600 samples injected into the cytosol of live A549 cells, where X = D10-t-DHLA, D6-p-DHLA, and D6-p-API. These images were acquired 30 min post-injection. The QDs dispersed throughout the cytosol but were excluded from the cell nuclei and other intracellular compartments. (B) Brightfield, PL, and overlay images of DHLA-QD600 samples after attempted microinjections (arrows) into A549 cells. These images were acquired 3 min post-injection. All scale bars = 25 µm. The pixel intensity calibration bars for the PL images are provided at the bottom. Data contributed by Hyungki Kim.   Potential cytotoxicity of (D10-t-DHLA)-QD600 on A549 cells was assessed using the 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium inner salt (MTS) viability assay. As shown in Figure 2.11, co-incubation of the cells with (extracellular) picomolar to micromolar concentrations of (D10-t-DHLA)-QD600 for either 3 h or 24 h did not result in any significant decrease in cell viability compared to control cells (not incubated with QDs). This result may reflect low cytotoxicity of the material, but is also consistent with the low 68  non-specific binding of (D10-t-DHLA)-QDs to cells and little or no spontaneous cellular uptake (see Figure 2.8), which may mitigate potential cytotoxicity (if any).      Figure 2.11 A549 cell viability assay with (D10-t-DHLA)-QD600 at concentrations between 10 pM and 1 μM. Cell viability is expressed as a percentage of the negative control (cells that were not incubated with QDs). Data points and error bars are the average and standard deviation of three replicates. The trendline is only to guide the eye.   2.2.6 Covalent conjugation and pH sensing  As a demonstration that dextran-functionalized QDs intrinsically support covalent conjugation strategies, (D6-t-DHLA)-QDs were reacted with fluorescein isothiocyanate (FITC), as illustrated in Figure 2.12A. Successful conjugation was confirmed by colocalization of QD and fluorescein PL on an agarose gel, with unreactive fluorescein as a control (see Figure 2.12B).    69    Figure 2.12 (A) Schematic for the reaction between dextran (as ligands on a QD, not shown) and FITC. (B) PL image of an agarose gel of FITC-labeled Dex-QDs and the PL emission spectra measured for each of the gel bands (labeled with Roman numerals).  The UV-visible absorption spectrum for the conjugates indicated an average of ~72 FITC molecules per QD. To leverage this conjugation in parallel with the colloidal stability offered by dextran ligands, a proof-of-concept ratiometric pH sensing experiment was done. Fluorescein is well-known to have pH-dependent PL intensity between pH 5 and pH 8, arising in large part from the transition between the monoanion and dianion forms (pKa ~6.4) and difference in quantum yield between the two forms.158 In contrast, QD PL was relatively insensitive to pH over the same range (see Figure 2.13).    Figure 2.13 Emission spectra for (D6-t-DHLA)-QD600 (left) and FITC (right) at various pH values. 70  Figure 2.14 shows that the fluorescein/QD600 emission ratio increased sigmoidally as the pH increased from pH 4 to pH 8, consistent with conversion to the dianion. The fluorescein and QDs were excited directly and concurrently.     Figure 2.14 Proof-of-concept pH sensing with FITC-(D6-t-DHLA)-QD600 (clockwise): (i) photograph of a well-plate loaded the FITC-labelled QDs in buffers at different pH; (ii) PL emission spectra of the FITC-(D6-t-DHLA)-QD600 at in buffers at various pH; (iii) plot of the FITC/QD600 PL emission ratio as a function of pH.  2.2.7 Immunolabeling with tetrameric antibody complex conjugates  A TAC is an antibody complex that is comprised of four antibodies. Two antibodies, which are specific to their target, are held together by two more identical antibodies, which bind to the Fc region of the antigen-specific antibodies (i.e. non-antigen binding domain).159   A model sandwich immunofluorescent assay for EPO was developed using Dex-QDs as a reporter and a TAC that is specific for both the dextran coating on the QDs, and the EPO analyte  (Figure 2.15). The assembly of TAC-EPO on the Dex-QDs was visualized using agarose gel electrophoresis and is marked by a cathodic shift upon TAC binding. A proof-of-concept immunoassay with the Dex-QD TAC-EPO bioconjugates showed a photoluminescent contrast ratio of 8:1 for samples consisting of 5 mU EPO and no EPO, with a p-value of 0.04, confirming the selectivity of the Dex-QD bioconjugates.  71    Figure 2.15 EPO immunofluorescent assay with Dex-QD TAC conjugates. (A) Cartoon schematics and agarose gels of Dex-QDs binding with (i) anti-dextran antibody and (ii) a full TAC. Binding is indicated by electrophoretic mobility shifts and increased streaking. (B) Proof-of-concept sandwich immunoassay for EPO: (i) cartoon schematic of the assay format with (D10-t-DHLA)-QD645 and (ii) measured PL intensity for a sample with EPO (5 mU) and without EPO. Error bars are the average of three replicates. The contrast ratio was 8:1, with a p-value of 0.04. Data contributed by Melissa Massey.  A second demonstration of the utility of a Dex-QD-TAC conjugate was immunolabeling of fixed SK-BR3 cells, which express HER2 on their surface. In this case, the TAC included an anti-dextran antibody and an anti-HER2 antibody. Figure 2.16 shows that the TAC-conjugated QDs selectively labelled the cells, and that QDs without TAC showed negligible non-specific binding. The PL intensity contrast ratio between TAC and no TAC was 36:1.      Figure 2.16 Cellular immunofluorescent labelling with TAC conjugates of Dex-QDs. (A) Immunolabeling of HER2-positive SK-BR3 breast cancer cells: (i) cartoon schematic of the experiment; (ii) representative images of cells incubated with (D6-p-DHLAm)-QD600 with and without TAC-Anti-HER2 (scale bar = 80 µm in main images, 20 µm in inset); and (iii) average PL intensity measured for cells with (D6-p-DHLAm)-QD600 with and without TAC-Anti-HER2 (error bars are the standard deviation from imaging 11 cells). Images were acquired with the same microscope, camera, and image processing settings. 72  2.2.8 Discussion  Dextran exhibits many properties that are desirable as a coating for nanoparticles for applications in bioanalysis and imaging. It is neutral in charge, hydrates to a high degree, and has steric bulk—all of which are ideal properties in a colloidally stable, low-fouling coating.142,160,161 Endotoxin/pyrogen-free preparations of dextran are also available commercially. Moreover, dextran is readily functionalized at its terminus and, in contrast to PEG, along its backbone. It is also poorly soluble in solvents such as ethanol, enabling the use of precipitation and washing steps for quick and easy purification after chemical modification with small molecules. We took advantage of all these properties in this study. Despite the popularity of utilizing dextran as a coating for iron oxide nanoparticles, only a few studies on the dextran functionalization of QDs have been reported to date.162–164 A possible reason is that native dextran and carboxydextran have relatively weak affinity for QDs, in contrast to their high affinity for iron oxide nanoparticles, necessitating the use of dextran derivatives that are not commercially available. As we discuss next, the few previous reports of dextran-functionalized QDs have taken different synthetic approaches than our approaches here and were more limited in their demonstration of the potential capabilities of dextran-functionalized QDs in bioanalysis and imaging.   With respect to synthesis, the most similar approach to our methods started from aminodextran and modified it with 2-iminothiolane (Traut’s reagent) to introduce pendant monothiol linkers.162 This approach has less stable binding per anchor than our use of dithiols, and the thiol groups introduced by 2-iminothiolane are potentially unstable.165,166 Other approaches have been based on grafting dextran onto QDs pre-functionalized with an aminosilanized silica shell163 or encapsulated within a polyacrylate shell.164 These methods yield significantly larger hydrodynamic diameters and the preparation of the QDs is much more intensive. Other studies have functionalized QDs with dextran in the form of electrostatically-bound, 100–200 nm diameter aggregates of thioalkyl acid-coated QDs, a dextran derivative (e.g. carboxymethyl dextran or aminodextran), and, in one case, a second polyion.167,168 These materials are not functionally equivalent to monodisperse QDs.   With respect to demonstration of capabilities for bioanalysis and imaging, the studies to date with monodisperse dextran-functionalized QDs have shown that dextran minimizes non-specific uptake 73  by cells, and enables the detection of lectins through aggregate formation.162–164 Something that has been notably lacking is the preparation of bioconjugates with dextran-functionalized QDs, which may be a factor in their limited scope of applications to date. Here, we have addressed this limitation in several ways with our Dex-QDs. Labeling with FITC represents a basic proof of concept for covalent conjugation of the Dex-QDs and is potentially extendable to heterobifunctional crosslinkers. Finally, conjugation with the TACs addresses many of the common challenges with the preparation of nanoparticle-antibody conjugates. It ensures that the anti-target antibody is optimally oriented for binding, it acts as a spacer between the QD and the anti-target binding site, and conjugation is rapid (minutes) and spontaneous. Crosslinking reagents are not required, competing hydrolysis is not a challenge, and the removal of unconjugated antibodies is not a necessary step, greatly increasing reproducibility. Our use of these various conjugation strategies to demonstrate pH sensing, assays of proteolytic activity, an in vitro sandwich immunoassay, and immunolabeling of cells has greatly expanded the utility of dextran-functionalized QDs but is far from limiting.  Our evaluation of various dextran ligands (i.e. terminal versus pendant modifications, different anchoring groups, different molecular weights) revealed routes through which the Dex-QDs may be rationally designed and optimized for applications. For example, in applications where brightness is paramount, the D-p-API ligands are arguably the best candidate because the imidazole anchoring group does not diminish PL quantum yield like a thiol anchoring group. At present, the tradeoff is poorer colloidal stability versus thiol anchoring groups, but this disadvantage, and the relatively high levels of protein adsorption may potentially be mitigated by higher degrees of modification of dextran with API. If colloidal stability and resistance to non-specific binding are critical, then the D10-t-DHLA ligand is likely the best choice. Our results suggest that D-t-DHLA binds to QDs in an orientation and at densities that yield conformations of dextran that are somewhere between a mushroom-like conformation and brush-like, forming a thick and robust coating. We make this inference based on the lower non-specific binding and analogy with PEG coatings. (It is thought that PEG densities that correspond to the mushroom-to-brush transition are required to resist protein adsorption.169) In turn, the D-p-DHLA ligands may be an option that provides sufficient colloidal stability and resistance to non-specific binding but with reduced steric hindrance versus D-t-DHLA ligands. This combination of properties may, for 74  example, be desired when polyhistidine-tagged biomolecules are to be conjugated. Variation of molecular weight offers further tuning of properties. Prospective synthetic modifications of the dextran (e.g. with a small molecule therapeutic or contrast agent) may also lead to a preference for either API or dithiol (or initially the dithiolane) anchoring groups because of their different chemical reactivity profiles. The synthetic routes to our dextran ligands are also readily adapted to commercially available carboxydextran or aminodextran, in case these materials are preferred over native dextran for certain applications.   Last but not least, it should also be noted that our dextran ligands have potential applications with other inorganic nanomaterials because of the relatively broad affinity of thiol and imidazole groups for metals. Although we have not presented the results here, we have successfully functionalized gold nanoparticles and InP/ZnS QDs with our dextran ligands. We anticipate applicability with any QD material that has a zinc or cadmium chalcogenide surface, as well as platinum and a variety of other metal nanoparticles.   2.3 Conclusion   We have developed a versatile set of dextran ligands for the functionalization of colloidal semiconductor QDs. The ligands varied in molecular weight and the use of a terminal dithiol anchoring group, pendant dithiol anchoring groups, or pendant imidazole groups to bind the dextran to the QDs. Although each dextran ligand had particular advantages and disadvantages relative to the others, the Dex-QDs collectively offered robust colloidal stability and low levels of non-specific binding with proteins and cells, had little or no effect on cell viability, and supported multiple strategies for conjugation. The latter included covalent coupling, and the preparation of immunoconjugates via TACs. The utility and advantages of the Dex-QDs and these conjugates was demonstrated through proof-of-concept pH sensing, an in vitro fluoroimmunoassay for EPO, and selective immunolabeling of breast cancer cells. Dex-QDs are promising materials for a variety of applications in bioanalysis and imaging.   2.4 Experimental section  75  2.4.1 Materials  2-propanol (≥99.5%), dimethyl sulfoxide (DMSO; ≥99.9%), ethylenediamine (≥99%), dichloromethane (DCM; ≥99.5%), sodium hydroxide (≥97%), hexamethylenediamine (HMDA; 98%), sodium cyanoborohydride (95%), sodium (meta)periodate (≥99%), hydrochloric acid (ACS reagent, 37%), L-glutathione reduced (≥98.8%), L-histidine (≥99%), methanol (≥99.8%), chloroform (≥99.8%), tetramethylammonium hydroxide solution (TMAH; 25 wt% in methanol), tris(2-carboxyethyl) phosphine hydrochloride (TCEP, ≥98%), agarose, fluorescein isothiocyanate isomer I (FITC; ≥90%), fluorescein, rhodamine B, 1-(3-aminopropyl)imidazole (API), ethanolamine (ACS reagent, >99%), 5,5’-dithiobis(2-nitrobenzoic acid) (DTNB), anthrone (ACS reagent, 97%), sulfuric acid (95–98%), and Concanavalin A (ConA) were from Sigma-Aldrich (Oakville, ON, Canada). Glucose was from BDH Chemicals (Toronto, ON, Canada). Potassium permanganate was from Anachemia Science (Richmond, BC, Canada). Ninhydrin monohydrate was from Amresco (Dallas, TX). 5-(1,2-dithiolan-3-yl)pentanoic acid (lipoic acid, LA) was from Oakwood Chemical (Estill, SC). N-hydroxysuccinimide (NHS), anhydrous sodium sulfate, and tetrahydrofuran (THF) were from EMD Millipore (Burlington, MA). 2,4,6-trinitrobenzene sulfonic acid (TNBS; 1% in MeOH) was from G-Biosciences (St. Louis, MO).   Water (sterile, nuclease free; defined as ultrapure H2O or UPH2O) and glycerol (sterile) were from VWR International (Mississauga, ON, Canada). Water was from a Milli-Q system (Millipore Sigma, Burlington, MA) and had a specific resistance ≥18 MΩ cm. Unless otherwise specified, water refers to that from the Milli-Q system. HEPES was from Sigma-Aldrich (Oakville, ON, Canada). Potassium carbonate anhydrous (ACS grade), sodium bicarbonate (ACS grade), sodium tetraborate decahydrate (ACS grade), and citric acid (ACS grade) were from Amresco (Dallas, TX). Sodium chloride (ACS grade), ethylenediamine tetraacetic acid disodium salt dehydrate (EDTA), magnesium chloride hexahydrate, potassium chloride, tris-borate-EDTA 10× solution, and sodium phosphate dibasic heptahydrate were from Fisher Scientific (Ottawa, ON, Canada). Potassium phosphate monobasic was from EMD Millipore (Burlington, MA). Calcium chloride dehydrate was from Anachemia Science (Richmond, BC, Canada). Buffers were filtered through a 0.22 μm-porous membrane filter prior to use.  Dextran from Leuconostoc mesenteroides (~6000 MW, lot number D-4967A) was from Dextran Products Limited (Scarborough, ON, Canada). Dextran from Leuconostoc spp. (Mr ~6000, 4500–76  7500 Da, product number 31388) and dextran from Leuconostoc mesenteroides (average mol wt. 9000–11000 Da, product number D9260) were from Sigma-Aldrich. Monoamine dextran (3.5, 6, and 10 kDa) were from Fina Biosolutions (Rockville, MD) and were used as a starting material in some cases. No differences in the resulting ligands were observed, and so the ligands are differentiated only by the dextran molecular weight and source bacterial strain. The (L) and (M) notations are used for Leuconostoc spp. and Leuconostoc mesenteroides, respectively. The purity of the commercial dextran was estimated by 1H NMR to be >99% carbohydrate. The < 1% of small-molecule impurity was removed during ethanol precipitation steps.   Bovine serum albumin (BSA) was from Amresco (Dallas, TX, USA). Lysozyme from chicken egg white and bovine plasma were from Sigma-Aldrich (Oakville, ON, Canada). Skim milk powder was from the local grocer.   TAC Anti-EPO and TAC Anti-HER2 complexes were prepared using the EasySepTM Do-It-Yourself Selection Kit (STEMCELL Technologies, Vancouver, BC, Canada). Anti-HER2 antibody (NBP2-32863) was from Novus Biologicals (Burlington, ON, Canada). Anti-EPO antibody (EPO-16, clone 16F1H11, mouse monoclonal antibody to human erythropoietin) was from STEMCELL Technologies. The erythropoietin (EPO) ELISA Kit and lyophilized human recombinant EPO (rhEPO) were from STEMCELL Technologies. rhEPO was reconstituted in sterile water (5 μg/25 μL) and diluted to 1.0 mL with Buffer B from the ELISA kit (PBS buffer with additives).  Chloroform-D (D 99.8%) was from Cambridge Isotope Laboratories (Tewksbury, MA). Deuterium oxide (99.9 atom% D) was from Sigma-Aldrich.  QDs were synthesized using standard hot-solvent methods170–175 and their concentrations were determined from the measured absorbance at their first exciton peak.176 QDs are denoted as QDλ, where λ is the peak PL emission wavelength. QD590 and QD600 were CdSe/ZnS core/shell materials and QD520, QD600, QD630, and QD645 were CdSe/CdS/ZnS core/shell/shell materials. 77  2.4.2 Instrumentation UV-visible absorption and PL spectra were acquired using an Infinite M1000 Pro plate reader (Tecan Ltd., Morrisville, NC). Monochromator bandwidths were typically 5–7 nm with excitation wavelengths of 400–450 nm depending on the experiment. Brightfield, differential interference contrast (DIC), and fluorescence imaging were done with an IX83 inverted epifluorescence microscope (Olympus, Richmond Hill, ON, Canada) equipped with an X-Cite 120XL metal-halide light source (Excelitas Technologies, Mississauga, ON, Canada), an Orca-Flash 4.0 V2 sCMOS camera (C11440; Hamamatsu Photonics, Hamamatsu, SZK, Japan), motorized filter wheels (Sutter Instruments, Novato, CA), and MetaMorph/MetaFluor software (Molecular Devices, Sunnyvale, CA). For cell immunolabeling and non-specific binding studies, the filter set was 405/20 (center line/bandwidth in nm) for the excitation filter; a 590 nm cut-off dichroic mirror; and a 600 nm longpass emission filter. Filters and dichroic mirrors were from Chroma Technology Corp (Bellow Falls, VT, USA). ImageJ software (NIH, Bethesda, MD) was used for processing images.  NMR spectra were collected with a Bruker AV III HD 400 MHz spectrometer (Bruker, Billerica, MA). IR spectra were acquired using a Frontier FT-IR spectrometer with attenuated total reflectance (ATR) sampling and a ZnSe ATR crystal, with data collected over the wavenumber range of 4000 to 650 cm–1 (PerkinElmer, Waltham, MA, USA). Unless otherwise noted, agarose gels were imaged using a Gel Doc XR+ System (Bio-Rad Laboratories Inc., Hercules, CA).   Dynamic light scattering (DLS) measurements were made on a Nanobrook Omni instrument (Brookhaven Instruments Inc., Long Island, NY). Nanoparticle tracking analysis (NTA) was done using a NanoSight NS300 (Malvern Panalytical Ltd., Malvern, Worcestershire, UK). The source laser was 488 nm and a 500 nm long-pass filter were used to block scattered laser light for fluorescence mode NTA measurements.   Capillary electrophoresis (CE) was performed using an Agilent 7100 CE system (Agilent Technologies, Saint Laurent, QC, Canada) equipped with hydrodynamic injection, a fused silica capillary with an inner diameter of 50 μm, and a diode array detector for UV-visible absorption. 78  The effective length of the capillary was 52 cm. The applied potential was 25 kV with a current of 60 µA.  2.4.3 Dextran ligands and QD functionalization  For terminal modification, the reducing end of dextran was functionalized via reductive amination with hexamethylenediamine (HMDA). The aminated dextran product was purified via ethanol precipitation, dried under vacuum, and reacted with lipoic acid-succinimidyl ester (LA-NHS) to yield dithiolane-functionalized dextran. Commercial monoamine dextran, if used, was analogously reacted with LA-NHS. For pendant modifications, dextran was partially oxidized with sodium (meta)periodate and one of three reaction pathways was followed: reductive amination with HMDA and subsequent reaction with LA-NHS; reductive amination with N-(2-aminoethyl)lipoamide; or reductive amination with API. Ligands were characterized by 1H NMR, FTIR, and colorimetric assays for functional groups.  Dex-QDs were prepared in two steps. Hydrophobic, organic-phase QDs were first transferred to the aqueous phase via ligand exchange with histidine.177 Dextran ligands were then incubated with the His-QDs at 60 °C for 3 h. Dex-QDs were collected and purified via spin filtration. Dithiolane-modified dextran ligands were reduced with TCEP prior to ligand exchange with His-QDs. Dex-QDs were characterized by dynamic light scattering (DLS; Nanobrook Omni, Brookhaven Instruments, Long Island, NY), nanoparticle tracking analysis (NTA; NanoSight NS300, Malvern Panalytical, Malvern, Worcestershire, UK), agarose gel electrophoresis, capillary electrophoresis (7100 system; Agilent Technologies, Saint Laurent, QC, Canada), and functional tests.   2.4.4 Functional tests 2.4.4.1 Anthrone assay with QDs A stock solution of dextran from Leuconostoc spp. (~6 kDa) in water was prepared (~2 mg/mL). This solution was used to prepare dextran standards ranging from 0–300 μg/mL. X-QD600 samples (X = DHLA, D10-t-DHLA, D6-t-DHLA (L), D6-p-API, D6-p-DHLA, and D6-p-DHLAm) were diluted 400-fold in water to concentrations of ~3.2–4.4 nM. An anthrone solution was prepared in concentrated sulfuric acid (2 mg/mL). Aliquots of the dextran standards and samples (125 μL) were added to microcentrifuge tubes. Anthrone solution (375 μL) was then added 79  to each of the microcentrifuge tubes. The tubes were carefully inverted and then cooled at 4 °C for 10 min. The tubes were inverted again to ensure complete mixing before incubation at ~80 °C for 45 min. The tubes were again cooled at 4 °C for 10 min. Aliquots of each of the solutions (100 μL) were transferred into a 96-well plate in quadruplicate. The absorbance was measured at 630 nm. A calibration curve of absorbance at 630 nm versus dextran concentration (μg/mL) was prepared and used to determine the quantity of dextran in each of the ligand samples. The number of dextran chains per QD was estimated from the quantity of dextran determined for each of the QD samples.  2.4.4.2 Tests with ConA, BSA, and lysozyme  All samples were prepared to a final concentration of 50 nM (D10-t-DHLA)-QD600. Briefly, the QDs were incubated with 100 equivalents BSA, lysozyme, or ConA which were all prepared in 1× PBS supplemented with 1 mM CaCl2 and MnCl2. The concentrations of the various proteins were 5 µM in all the samples prepared. A control sample was prepared with no protein added. The samples were spiked with glycerol to a final concentration of 2% (v/v). An agarose gel was run for 30 min at ~6.7 V cm–1 in 1× TBE buffer (100 mM, pH 8.3) and imaged under UV illumination.  2.4.5 Electrophoretic mobility 2.4.5.1 Agarose gel electrophoresis Unless otherwise stated, agarose gels (1.0% w/v) were prepared in 1× TBE buffer (pH 8.3) and QD samples (~1 pmol) were diluted with bicarbonate buffer (100 mM, pH 9.3) and spiked with 50% v/v glycerol solution (1–2 μL, final glycerol content 10% v/v for the sample). The gels were run for ~30 min at ~6.7 V cm–1 and imaged under UV illumination.  2.4.5.2 Capillary electrophoresis  For CE experiments, the capillary was preconditioned by rinsing in sequence with methanol, 0.1 M NaOH (aq), water, and borate buffer (20 mM, pH 9.3) for 5 min each. All samples were prepared in borate buffer with added fluorescein (100 µM) as an internal standard. QD600 samples were 200 nM. Rhodamine B (100 µM) was used as a reference neutral analyte. The capillary was rinsed with borate buffer (20 mM, pH 9.3) for 4 min prior to each injection (5 s injection time, 50 mbar injection pressure). Borate buffer (20 mM, pH 9.3) was used for the sample runs. QDs were hydrodynamically injected at the anodic side of the capillary and traveled towards the cathode with 80  the electroosmotic flow. Following each sample run, the capillary was post-conditioned by successive rinses with 0.1 M NaOH and water for 3 min each. The pressure was 1 bar during rinses and the sample runs. The temperature was maintained at 20 °C. Data were recorded at multiple wavelengths, but the electropherograms were plotted for the signal at 260 nm. All CE runs were performed in triplicate.   2.4.6 Cell-based methods  A549 human lung carcinoma cells, MDA-MB-231 human epithelial breast cancer cells, and SK-BR3 human breast cancer cells were from ATCC (Manassas, VA) and cultured in growth media under standard conditions (5% CO2 at 37 °C). Where applicable, cells were fixed with paraformaldehyde. Cell viability was assessed using a commercial MTS assay kit (Abcam, Toronto, ON, Canada).   2.4.6.1 Cell viability The potential cytotoxicity of (D10-t-DHLA)-QD600 was assessed using a 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium inner salt (MTS) assay kit (Abcam, Toronto, ON, Canada). A549 cells were seeded in a 96-well tissue culture-treated clear-bottom plate (ThermoFisher, Waltham, MA) at ~10 000 cells/well and grown overnight in a humidified incubator with 95% air/5% CO2 at 37 °C. Cells were washed with HEPES-KRH buffer (100 µL; pH 7.2, 120 mM NaCl, 5 mM KCl, 2mM CaCl2, 1 mM MgCl2, 25 mM NaHCO3, 5.5 mM HEPES, 1 mM D-glucose) and incubated with QD solutions (10 pM–1 µM) for either 3 h or 24 h in a humidified incubator with 95% air/5% CO2 at 37 °C. Following incubation, the QD solutions were removed and the cells were washed with HEPES-KRH buffer (100 µL). McCoy’s 5A media without phenol red (90 µL) was then added to the cells, which were left to grow for 3 days in a humidified incubator with 95% air/5% CO2 at 37 °C. After this proliferation period, MTS reagent solution (10 µL) was added to each well and the cells were incubated for 2 h in a humidified incubator with 95% air/5% CO2 at 37 °C. The absorbance was measured at 490 nm (analytical wavelength) and 650 nm (for background subtraction). Cellular viabilities were reported as the percentage of the absorbance for negative control wells (non-treated cells). The assays were performed in triplicate. 81  2.4.6.2 Non-specific binding studies with cells  MDA-MB-231 and A549 cells were grown in tissue-culture treated, cover-glass bottom, 8-well chamber slides (Eppendorf, Hamburg, Germany). Each well was seeded with ~30 000 cells. The cells were grown until confluency (between 72–96 h) before proceeding with non-specific binding experiments.   Once cells were confluent, the media was removed and the cells were washed with HEPES-KRH buffer (300 µL). X-QD600, where X = GSH, D6-t-DHLA (L), D10-t-DHLA, D6/10-t-DHLA, D6-t-DHLA (M), D6-p-DHLA, and D6-p-API, were prepared at a final concentration of 50 nM in HEPES-KRH buffer. (D6-p-DHLA and D6-p-API were prepared from D6 (L).) The washed cells were incubated with the QD solutions for 30 min in a humidified incubator with 95% air/5% CO2 at 37 °C. The QD solutions were removed, and the cells were washed twice with HEPES-KRH buffer (300 µL) to remove unbound QDs. The samples were then imaged in 300 µL HEPES-KRH buffer. The assays were performed in triplicate. Imaging parameters are described in section 2.4.2.  A quantitative assay was also performed with A549 and MDA-MB-231 cells to evaluate the non-specific binding of X-QD600, where X = GSH, D10-t-DHLA, D6-p-DHLA. Cells were seeded into a 96-well tissue culture-treated clear-bottom plate (ThermoFisher, Waltham, MA). The cell media was removed, and the cells were washed with HEPES-KRH buffer (100 μL). Solutions of QDs were prepared in HEPES-KRH buffer at concentrations of 10, 30, 50, 70 and 100 nM. Aliquots of the QD samples (90 μL) were added to the cells and the cells were incubated for 30 min in a humidified incubator with 95% air/5% CO2 at 37 °C. The QDs were removed and the cells were washed twice with HEPES-KRH buffer (100 μL) before absorbance and PL emission spectra were collected on a plate reader. The assay was performed in duplicate. Acquisition parameters are described in section 2.4.2.  2.4.7 Applications  2.4.7.1 Covalent conjugation and pH sensing FITC-labeling of (D6-t-DHLA)-QDs FITC-labeled (D6-t-DHLA)-QD600 were prepared by mixing an aliquot of QDs (7.5 µL, 1 µM) in bicarbonate buffer (0.1 M, pH 8.8) with a FITC solution (0.5 mg/mL, 7.5 µL) in DMSO. The 82  reaction was left overnight at room temperature and the QDs were purified from unreacted FITC using a 10 kDa MWCO spin-filter (VWR International, Mississauga, ON, Canada). A negative control sample was prepared analogously but used fluorescein (no reactive group) instead of FITC. The labeled QDs were diluted to a final concentration of ~0.5 µM in bicarbonate buffer.   Additional control samples were prepared with only FITC or fluorescein (no QDs) by mixing an aliquot (0.5 mg/mL in DMSO, 7.5 µL) with bicarbonate buffer (100 mM, pH 9, 7.5 µL). A control sample of (D6-t-DHLA)-QD600 was prepared by diluting an aliquot of QDs (7.5 µL, 1 µM) with bicarbonate buffer (100 mM, pH 9, 7.5 µL). Samples were loaded into a 1% w/v agarose gel prepared in 1× TBE buffer, and run at a field strength of ~6.7 V cm–1 for 12 min. The gel was imaged under UV illumination.   pH Sensing To prepare pH sensors, the above reaction was scaled up. (D6-t-DHLA)-QD600 were diluted to a final concentration of 1 µM in bicarbonate buffer (100 µL, 100 mM, pH 9) and mixed with FITC (0.5 mg/mL in DMSO, 100 µL). The sample was left overnight at room temperature, protected from light, with shaking. To separate labeled QDs from excess dye, the QDs were precipitated with absolute ethanol (400 µL). The labeled QDs were pelleted via centrifugation, and the supernatant (containing unreacted FITC) was removed. The QDs were resuspended in UPH2O, and the precipitation and washing step was repeated once more. The QD pellet was dried under vacuum and then resuspended in UPH2O (200 μL) at a final QD concentration of ~0.5 µM.  For pH sensing experiments, FITC-labeled QD samples (20 μL, 0.5 µM) were mixed with an equal volume (20 μL) of the following buffers: 100 mM MES buffers (pH 4.5 and 6.5), 100 mM sodium phosphate buffer (pH 7.0), 100 mM borate buffer (pH 8.5), and 100 mM bicarbonate buffers (pH 8.5, 9.1 and 10.5). The samples were transferred to a 96-well plate and PL emission spectra (475–800 nm, 2 nm step-size, 450 nm excitation wavelength) were measured.  2.4.7.2 Conjugation with TAC and applications Preparation of TAC anti-target immunocomplexes 83  Bifunctional anti-target TACs, which consisted of an anti-dextran antibody on one end and an anti-target (e.g. EPO, HER2) binding antibody on the other end were prepared according to the instructions of an EasySepTM Human “Do-It-Yourself” Positive Selection Kit II immunomagnetic positive selection cell isolation kit (STEMCELL Technologies, Vancouver BC, Canada). The anti-target antibody was either a mouse monoclonal anti-human EPO antibody (EPO-16, clone 16F1H11, mouse monoclonal antibody to human erythropoietin; STEMCELL Technologies) or a mouse anti-human HER2 antibody (clone HRB2/282; Novus Biologicals, Burlington, ON, Canada). The desired anti-target antibody (15 μg) was mixed with kit Component A (100 μL) and kit Component B (100 μL) solutions, in sequence. The sample was incubated overnight at 37 °C then diluted with 1× PBS buffer (up to 1 mL). TACs were stored at 4 °C until needed.   SK-BR3 immunolabeling and imaging A suspension of fixed SK-BR3 cells (in 1× PBS) were immunolabeled with TAC anti-HER2 and (D6-p-DHLAm)-QD600. SK-BR3 cells (10 µL, 0.5 × 106 cells/mL) were pipetted into a 1.7 mL Eppendorf tube, followed by 12.3 µL of TAC anti-HER2 (81 nM, or 15 µg/mL, 1 pmol), and, lastly, (D6-p-DHLAm)-QD600 (0.5 pmol in water). A control sample, which did not include TAC anti-HER2 complex, was prepared as above but was spiked with 12.3 µL of 1× PBS buffer instead. The samples were mixed via pipette and incubated at room temperature, protected from light, for 30 min. The samples were pelleted via centrifugation at 55 RCF for 5 min. The supernatant, which contained excess/unbound QDs, was removed via pipette. The cells were then resuspended in 20 µL of 1× PBS. For imaging, the samples were drop cast on a microscope slide and a cover slip was applied. The samples were inverted and imaged through the cover slip. The cell imaging parameters are described above in section 2.4.2.  84   Development of single-chain polymer nanoparticles for cellular immunolabeling   This chapter is an adaptation of Bajj, D.N.F.; Tran, M.V.; Tsai, H.-Y.; Kim, H.; Paisley, N.R.; Algar, W.R.; Hudson, Z.M., Fluorescent Heterotelechelic Single-Chain Polymer Nanoparticles: Synthesis, Spectroscopy, and Cellular Imaging. ACS Appl. Nano Mater. 2019, 2, 898–909, with permission from the American Chemical Society (Copyright 2019). Refer to the Preface for full details of author contributions. Unless indicated in figure captions, I had a sole or significant role in obtaining the data presented in this chapter.   3.1 Introduction  Many cancer diagnostic tests characterize differences in the antigenic expression of intracellular and extracellular proteins between different forms of malignant cells. In breast cancers, the most common antigens that are investigated are the estrogen receptor (ER), progesterone receptor (PR), and HER2.178 The ER and PR antigens are intracellular, situated within the nuclear membrane, whereas the HER2 antigen has a portion of its structure in the extracellular domain. Quantification of these antigens is important for determining the correct treatment for a patient, including immunotherapy, which aims to boost the body’s immune system response towards these cancerous cells. Other types of human cells that are analyzed using immunofluorescence include stem cells, erythrocytes, and leukocytes.179–181 Furthermore, immunofluorescent labeling is not only limited to human cells, but also pathogenic cells such as bacterium,70,130 and viruses.126,182   There are several techniques that allow researchers to qualitatively or quantitatively assess the antigen expression levels of cancer cells, including immunohistochemistry183 and flow cytometry.184 These techniques utilize target-specific antibodies that are labeled with fluorescent reporters. The most commonly utilized fluorescent reporters are organic dyes such as fluorescein and the Alexa Fluor series dyes. The advantage of using organic dyes is that their relatively small size allows investigating cancer cells without perturbation of antigen-antibody binding.185 These organic dye labels, however, typically lack photostability, and their fluorescent signal will degrade 85  under continuous or intense illumination, significantly hindering their usefulness for antigen quantification, especially for antigens of low abundance. Furthermore, self-quenching of the fluorescent dyes is an issue when multiple dyes are conjugated to the antibody, therefore, precise control over the stoichiometry is required.185  Luminescent nanomaterials have garnered significant interest for use in bioanalytical and imaging applications as they may have better brightness, photostability, and other advantageous properties versus organic dyes.26,28,30,38,185,186   SCPNs, as depicted in Figure 3.1, are a diverse type of NP, which comprise a single-polymer chain that collapses either via covalent cross-linking, metal-coordination, or supramolecular chemistry.97,98 A single-polymer chain can be decorated with many fluorescent dyes within a single particle vector, thus providing additional benefits in comparison to using single fluorescent dyes. These benefits include higher relative brightness, and greater photostability. A common approach for preparing fluorescent SCPNs includes encapsulation of the fluorescent dye during the nanoparticle formation step.187 The fluorescent dye can also be covalently coupled to the single-chain polymer backbone with controlled stoichiometry. The latter approach is more desirable as it is less prone to fluorescent dye leaching and was therefore utilized in our SCPNs. In comparison to other fluorescent NP materials, SCPNs are relatively new materials when applied in the field of bioimaging. Novel SCPN architectures and bioconjugation strategies are being investigated. Biological applications with SCPNs are diverse and include drug delivery, magnetic resonance imaging, and intracellular studies. Recent work by Albertazzi and Palmans demonstrated the preparation of biocompatible SCPNs that were nontoxic towards cells, and these SCPNs enabled spatially controlled cell death via light-harvesting porphyrin groups that generated reactive singlet-oxygen species.187 Like these authors, we reasoned that SCPNs held promise for immunofluorescent cellular labeling applications, potentially combining moieties for targeting, and imaging. In this chapter, the development of fluorescent SCPNs for specific, extracellular labeling of the HER2 antigen on SK-BR3 breast cancer cells is demonstrated.    86    Figure 3.1 Cartoon schematic of SCPNs. Approximate hydrodynamic diameter: 26 nm.  To our knowledge, extracellular labeling with fluorescent SCPNs has yet to be achieved. Furthermore, there have been no applications of SCPNs that have leveraged antibodies for cellular targeting. As proof-of-concept with our SCPNs, we terminated the polymers with a biotin functional group, thus providing the ability to target cells via biotin-NeutrAvidin bridge. The SCPN-mediated cellular labeling strategy is shown in Figure 3.2.   87    Figure 3.2 Strategy for immunolabeling fixed SK-BR3 cells with SCPN-Biotin.   3.2 Results and discussion  3.2.1 Synthesis of SCPNs The SCPNs were prepared by the Hudson research group by using a poly(pentafluorophenyl) acrylate (pPFPA) polymer as a SCPN precursor. These polymers were prepared via reversible addition-fragmentation chain-transfer (RAFT) polymerization of PFPA monomers. The reactive ester backbone of the polymer was reacted with 10% amine-functionalized chiral benzene-tricarboxamide (BTA-NH2), which promotes chain collapse and formation of the SCPNs. A further 10% was reacted with fluorescent amine-modified FITC. The remaining 80% of the polymer backbone was functionalized with an amine-terminated polyether (Jeffamine M-1000), as to facilitate aqueous solubility and biocompatibility for cellular immunolabeling studies. Final polymer structure (P5) is shown in Figure 3.3. Upon introduction to water or aqueous buffers (e.g. borate, phosphate, acetate), the P5 polymer folded and was readily dispersible as a SCPN. The P4 88  and P5 SCPNs were analyzed via DLS, where the mass-weighted DLS distribution showed a mean hydrodynamic diameter of 12 nm and 26 nm, respectively.   Figure 3.3 SCPN-forming polyacrylamides. The polymer chain collapsing agent (BTA-NH2) is in orange. The fluorescent dye (FITC) is in green. P5 is the precursor polymer to P6. The cell-labeling bioconjugate handle (biotin) for P6 and cell-targeting molecule (FA) for P7 are in pink. The cytotoxic payload (CPT) for P4 is in red. Daniel Bajj synthesized the polymers.    SCPNs with an additional functional group at the polymer backbone terminus were also prepared. Biotin-functionalized SCPNs (Figure 3.3, SCPN-Biotin, P6) were prepared by copper-catalyzed azide-alkyne bioconjugation between the azide moiety on the P5 and alkyne-functionalized biotin. Folic acid (FA) conjugated SCPNs (Figure 3.3, CPT-SCPN-FA, P7) were prepared analogously starting from SCPN-P4 and alkyne-functionalized folic acid.  89  The SCPN-P7 were utilized in photophysical characterization experiments and cellular viability assays, and SCPN-P6 was employed in specific cellular immunolabeling of SK-BR3 breast cancer cells.  3.2.2 Photophysical characterization of SCPNs 3.2.2.1 Photophysical properties In order to evaluate fluorescent SCPNs as a material for cellular labelling, an understanding of their photophysical properties (i.e. of dye molecules in the SCPN microenvironment) was first required. Selected photophysical properties of SCPN-P7 were compared to fluorescein, FITC, and FITC-labeled bovine serum albumin (BSA-FITC) as reference materials (Table 3.1). The BSA-FITC sample was selected as a model for FITC conjugated to amine groups on a macromolecule within defined structure and folding, albeit that the folding analogy is imperfect (the BSA is folded prior to labeling; the SCPN is folded after labeling). Nevertheless, the BSA-FITC is a useful reference because the properties of FITC (and other fluorescein derivatives) change between folded and unfolded BSA.188,189  Table 3.1 Photophysical properties of SCPN–P7 and reference materials. Material Φ a τ (ns) b r  c kPB (min–1) d kq (109 M–1 s–1) e Fluorescein 0.81 4.0 0.024 0.17 ± 0.02 2.84 ± 0.46 FITC 0.60 3.9 0.025 0.10 ± 0.01 2.68 ± 0.11 BSA-FITC 0.28 1.0 (14%) 3.8 (86%) 0.191 0.29 ± 0.04 (32%)  0.021 ± 0.002 (68%) ~0 SCPN-P7 0.12 1.8 (27%) 4.3 (73%) 0.201 0.3 ± 0.2 (19%) 0.024 ± 0.004 (81%) 0.43 ± 0.27 a Fluorescence quantum yield. All values ± 5%. b Fluorescence lifetime(s). All values ± 0.1 ns. c Fluorescence anisotropy. All values ± 0.001. d Photobleaching rate(s). e Stern-Volmer quenching constant with iodide.   3.2.2.1.1 Spectra and quantum yield Figure 3.4A shows that fluorescein, FITC, and FITC-BSA had approximately the same absorption, fluorescence excitation, and fluorescence emission band shapes, except for small (≤ 10 nm) spectral shifts between them. These SCPN-P7 exhibited several spectral features distinct 90  from the reference materials. In particular, the absorption spectrum of SCPN-P7 had a very pronounced hypsochromic shoulder, a similar but much less pronounced feature was present in its excitation spectrum, and its emission spectrum was more spectrally broad, suggesting a weak bathochromic spectral feature. This data cumulatively suggested the presence of multiple dye species and/or environments in the SCPN.   The main dye species was simply conjugated FITC with spectral properties similar to the reference materials. The reference material spectra were dominated by the properties of the dye dianion, which is the brightest form of fluorescein and generally dominant at pH ≥ 7. One possible secondary dye species within the SCPN-P7 appeared to be the dye monoanion, as suggested by the shoulder in the fluorescence excitation spectrum († in Figure 3.4A, ii) and broadening of the emission spectrum (‡ in Figure 3.4A, iii). Although inner filter effects (e.g. in the form of spectral shifts) can occur with the SCPN samples, we confirmed that these do not account for the increased width of the emission spectrum nor the shoulder in the excitation spectrum. The monoanion is normally only observed at pH < 6, but has an absorption peak and shoulder at shorter wavelengths than the dianion (~470 nm and ~450 nm, respectively, versus a peak at ~490 nm for the dianion), and an emission peak and shoulder at longer wavelengths (~510 nm and ~550 nm, respectively, versus a peak at ~515 nm for the dianion), which may account for the new spectral features († and ‡).158 An alternative interpretation is that the new features in the fluorescence spectra represent a sub-population of dianion that is in an environment different than the main population, as fluorescein is known to exhibit spectral shifts in response to changes in polarity and hydrogen-bonding interactions;190,191 however, the absorbance, excitation, and emission spectra usually shift in the same direction, which is not consistent with the hypsochromic shoulder in the excitation spectrum and bathochromic broadening of the emission spectrum. If not from the monoanion, these features would arise from multiple emitting species, multiple local environments in the SCPN, or some combination thereof.  The other secondary dye species in the SCPN-P7 appeared to be a non-fluorescent or weakly fluorescent H-like dimer or aggregate, as indicated by the much larger hypsochromic shoulder  (* in Figure 3.4A, i) in the absorption spectrum versus the excitation spectrum. Fluorescein dimers and trimers have been reported to have peaks in their absorption spectra at ca. 457, 470, and 505 91  nm,192 which is consistent with the hypsochromic shoulder (*). The hypsochromic shoulder (*) was unaffected by dilution of the SCPN between 0.25–75 µM, supporting an intra-particle H-dimer rather than an inter-particle interaction.  The SCPN quantum yield (0.12) was about one fifth the measured value for FITC. The fluorescein monoanion is reported to have a molar absorption coefficient and quantum yield that are ~2.5-fold smaller than the corresponding values for the dianion.158 The presence of a minority sub-population of monoanion may account for some of the apparent decrease in quantum yield, but not all of it, which further supports the formation of some intra-particle H-dimers or trimers.   3.2.2.1.2 Lifetime Figure 3.4B shows that SCPN-P7 had a biexponential fluorescence decay with a slow lifetime component (~4.3 ns) that was similar to the lifetime for fluorescein and FITC alone. We attribute this component to the dianion, where the combination of low quantum yield and largely unchanged lifetime was consistent with non-fluorescent H-dimers within the SCPN. The fast lifetime component (~1.8 ns) of the SCPN-P7 fluorescence decay may be due to the presence of fluorescein monoanion. A fluorescence lifetime of 3.0 ns is expected from the monoanion, but the apparent elevation in its pKa also suggests that a change in its lifetime from the local environment of the SCPN is possible. Alternatively, the short lifetime component could reflect that a fraction of dianions (and perhaps monoanions, if present) engage in some form of energy transfer to sinks within the SCPN, or that a sub-population of dianion is in an environment that causes a faster non-radiative relaxation rate.   Given that the SCPN and reference materials were measured under the same conditions, the possible presence of dye monoanion and probable presence of H-like aggregates is attributable to folding of the SCPN and its degree of labeling. Plane-to-plane stacking of the FITC molecules may be a passive side effect of the interactions between the BTA, and/or the FITC may actively engage in hydrogen bonding with the BTA. The latter is consistent with the possible presence of dye monoanion, as hydrogen bonding interactions with the FITC phenol group may elevate the phenol pKa from its usual value of 6.0–6.4.  92  3.2.2.2 Additional fluorescence measurements Other evidence for inclusion of FITC within the folded interior of the SCPN came from fluorescence anisotropy, photobleaching, quenching, and unfolding measurements. The fluorescence anisotropy measured from SCPN-P7 was approximately an order of magnitude larger than the anisotropy of the fluorescein and FITC (see Table 3.1), indicating a significant loss of rotational freedom within the nanoparticle. Figure 3.4C shows that the majority of fluorescent dye associated with the SCPN-P7 bleached at least four-fold more slowly than the fluorescein and FITC, suggesting protection within the interior of the SCPN. A minority of fluorescent dye in the SCPN-P7 bleached at close to the same rate as the fluorescein and FITC, presumably representing a sub-population of dyes at or near the surface of the folded nanoparticle. These results are consistent with recent work by Zimmerman and coworkers, who have demonstrated that incorporation of fluorescent dyes into polymer particles can greatly improve their stability under continuous irradiation.193 Collisional quenching experiments with iodide ion were also consistent with protection of a majority of fluorescent dye molecules within the SCPN-P7 interior. The bimolecular quenching rate constants derived from the Stern-Volmer plots in Figure 3.4D were approximately six-fold smaller with the SCPN-P7 than with fluorescein and FITC. Overall, the photophysical properties of FITC associated with SCPN-P7 were most analogous to BSA-FITC, which is also a folded structure.     93    Figure 3.4 Photophysical characterization of SCPN-P7 and reference materials. (A) Spectra: (i) absorption; (ii) fluorescence excitation; and (iii) fluorescence emission. (B) Fluorescence decays. (C) Photobleaching curves. (D) Stern-Volmer plots for collisional quenching by iodide ion. All measurements were made in 1 PBS buffer at pH 7.2 (see section 3.4.1 for recipe) between 21–26 C. Error bars represent the standard deviation of three replicate measurements. The concentrations for each sample can be found in the section 3.4. Data collected and analyzed in collaboration with Hsin-Yun Tsai and Hyungki Kim.    3.2.3 Cell viability assay CPT-SCPN-N3 (P4) or CPT-SCPN-FA (P7) were dosed to SK-BR3 cells, which are known to overexpress folate receptors and were thus chosen as a model cell type for viability assays. For both materials, the cell viability was approximately constant across SCPN concentrations ranging from picomolar to micromolar (Figure 3.5). This trend, which reflected a lack of acute cytotoxicity, suggested that (i) the SCPN was quite benign to the cells, and (ii) that the SCPN-conjugated camptothecin was unavailable to exert its biological effects, or otherwise had greatly reduced potency. Typical IC50 values for free camptothecin are in the range of 679 ± 92 nM.194 Release of the camptothecin from the SCPN was hypothetically possible through acidic or 94  enzymatic hydrolysis of the ester linkage after uptake via the endolysosomal system, as implemented with other nanoparticle materials.195–197 There are several possible reasons that we saw no acute cytotoxic effect. The first possibility is that the folate-conjugated SCPNs were not efficiently taken up by the cells, which precluded the drug from exerting an effect. Inefficient uptake may have resulted if the folate remained tightly associated with or sterically occluded by the folded SCPN, or if the display of a single folate did not meet the threshold for ligand display to induce uptake at the SCPN concentrations tested. An example of the latter is the observation that a minimum number of cell-penetrating peptides (~25) need to be conjugated to a semiconductor QD to induce uptake.198,199 QDs are similar in size to our SCPN, and so the display of multiple folate molecules may have been necessary.   Another possibility is that there was indeed uptake of the SCPN, but that it was not observed by fluorescence microscopy because of rapid loss of the fluorescence signal in the endolysosomal system. Full degradation of the SCPNs should have resulted in release of the CPT drug and a cytotoxic effect, and thus was unlikely. However, if only the fluorescent dye component of the SCPN was degraded (e.g. by rapid photobleaching), then another possibility for the absence of toxicity is that the hydrophobic CPT was tightly associated with or sterically occluded by the folded SCPN, such that its release by esterase activity was inefficient. If that was the case, then the next-generation design of the SCPN should also include a mechanism for the SCPNs to unfold in the endolysosomal system. Accepting that the camptothecin was unable to exert a biological effect, our results are consistent with other studies that have little or no acute cytotoxicity for SCPNs,200 including other SCPNs with BTA and PEG side chains,201 and PEGylated SCPNs based on other folding motifs202,203 or crosslinking chemistries.103  95   Figure 3.5 SK-BR3 cell viability assay with CPT-SCPN-N3 (P4), and CPT-SCPN-FA (P7). The nanoparticle concentrations were from 10 pM–16 µM (810–7–1.3 mg/mL) in PBS buffer (see Section 3.4 for recipe) and incubated with cells at 37 C. Cell viability is expressed as a percentage of the negative control (cells that were not incubated with the SCPN). Data points and error bars are the average and standard deviation of three replicates. The dashed lines are to guide the eye. The solid trendlines show that viability does not decrease as the concentration of SCPN increases.   3.2.4 Immunolabeling with SCPN-Biotin The biotinylated SCPN-P6 was used as a fluorescent probe for labeling SK-BR3 cells in a sandwich binding format with NeutrAvidin and biotinylated anti-HER2 antibody. The samples with all three components of the sandwich displayed significantly more fluorescence intensity versus the negative controls (Figure 3.6), demonstrating that fluorescent SCPNs are a viable tool for selective cell labelling. We note, however, that the magnitude of the contrast versus the negative controls was lower when older stock solutions of SCPN were used (see Figure 3.7 for an example using the same stock solution of SCPN-P6, obtained several weeks after that shown in Figure 3.6). Further studies will be needed to elucidate if this variation was inherent to the position of the biotin label on the polymer, inherent to the cultured cells, or whether the SCPNs experienced some degradation in aqueous solution over time.  96    Figure 3.6 Fluorescent labeling of SK-BR3 breast cancer cells with SCPN-Biotin (P6). Top row: brightfield images. Middle row: fluorescence images (the pixel intensity calibration scale is indicated on the right). Bottom row: merged brightfield and fluorescence images. The sample is in the right-most column. The “+” indicates that the cells were incubated with the respective material. Scale bars = 100 µm.  97    Figure 3.7 Fluorescent labeling of SK-BR3 breast cancer cells with SCPN-Biotin (P6). Top row: brightfield images. Middle row: fluorescence images. Bottom row: merged brightfield and fluorescence images. The sample is in the left-most column and negative controls are in the other three columns. The “+” indicates that the cells were incubated with the respective material. Scale bars = 100 µm. This figure represents a lower contrast example of the experiment shown in Figure 3.6, performed several weeks later using the same stock solution.   The SCPNs are a promising material for cellular immunolabeling applications. The diversity in the number of amine-conjugated dyes makes it possible to develop different colors of SCPNs, possibly making it feasible for multicolor immunolabeling applications. Despite its lower brightness in comparison to the other fluorescent nanoparticles evaluated in this thesis, there is significant capacity for improvement via different types of conjugated dyes, backbone polymers, or ligands for rendering hydrophilicity. The avidity of the SCPNs for the cellular antigenic targets can be improved by conjugating more biotin moieties per nanoparticle. The SCPNs demonstrated here were terminated with a single biotin molecule.   2132 (a.u.) 1638 (a.u.) Brightfield Fluorescence + SCPN-Biotin + NeutrAvidin + Biotin-AntiHER2  + SCPN-Biotin + NeutrAvidin + SCPN-Biotin Cells only Negative Merged 98  3.3 Conclusion   In summary, a strategy has been developed for the α and ω chain-end functionalization of SCPNs to give SCPNs suitable for bioconjugation. Fluorescein dyes in the SCPN microenvironment were found to be more resistant to photobleaching than free dye, experienced excitation and emission broadening, and were less bright because of effects associated with folding of the polymer chain. The SCPNs were conjugated with either biotin or folate, and biotin-functionalized SCPN-P6 was used successfully in the selective immunolabeling of SK-BR3 cells. SCPN-P7 had essentially no impact on cell viability when present in the extracellular environment at concentrations up to 16 µM, and minimal uptake of the SCPNs by endocytosis was observed. This work demonstrates that SCPNs are a viable platform for immunoconjugation and cell labeling, opening the door to the development of SCPNs for a broader scope of targets. Future work will look to improve the ease and robustness of the end group functionalization reactions, rationally design enhancements to the SCPN structure to benefit bioimaging (e.g. uptake by receptor-mediated endocytosis, better retention of native dye fluorescence properties), and optimize loading and release for SCPNs with therapeutic and contrast agent cargoes for applications in theranostics.  3.4 Experimental section  3.4.1 Materials Bovine Serum Albumin (BSA) was from Amresco. PBS buffers were from Gibco Life Technologies. The primary PBS buffer composition was pH 7.2, 1.54 mM KH2PO4, 2.71 mM Na2HPO4, 155 mM NaCl, and is denoted as 1 PBS. For samples with adhered cells, the PBS buffer composition was pH 7.2, 1.47 mM KH2PO4, 8.06 mM Na2HPO4, 138 mM NaCl, 2.67 mM KCl, 0.49 mM MgCl2, 0.90 mM CaCl2 and is denoted as PBS(+, +). All other reagents and solvents were from Sigma-Aldrich and used as received. Dialysis was carried out using a 3.5 kDa MWCO Spectra/Por 3 standard regenerated cellulose membrane. Details for the synthesis of the SCPNs are provided in the publication.204 The P4 and P5 polymers were prepared from azide-functionalized precursor PFPA polymers with Mn = 19 000 and 37 600, PDI = 1.34 and 1.38, respectively. 99  3.4.1.1 Preparation of FITC-labeled bovine serum albumin (BSA-FITC) A BSA solution was prepared by dissolving 39 mg (0.59 µmol) in 4 mL of 50 mM borate buffer pH 9.3 in a glass vial. Next, 235 µL of FITC solution (2.5 mM in 50 mM borate buffer pH 9.3) was added to the BSA solution. The sample was gently mixed in the dark at room temperature for 2 h. The labeled protein was dialyzed in 3.5 kDa MWCO dialysis tubing over two days against ~1 L of water at 4 °C with one water change to remove unreacted FITC. The calculated labelling ratio after purification was ~0.1:1 FITC:BSA.  3.4.2 Spectral characterization Absorbance and fluorescence excitation, emission, and anisotropy measurements were made with an Infinite M1000 Pro multifunction plate reader (Tecan, Morrisville, NC, USA) using 100 µL aliquots of samples and half-area 96-well UV-transparent plates (Corning, Kennebunk, ME, USA). The path length for absorbance measurements was 0.6 cm. A defined path length was not applicable to fluorescence measurements because an optical fiber bundle was used for both delivery of excitation light and collection of fluorescence emission. Sample concentrations were 10 μM for fluorescein and FITC, 1.0 μM for BSA-FITC (with respect to FITC concentration), and 75 μM for SCPN-P7. The step size was 1 nm for collecting spectra, the bandwidth was 5 nm, 430 nm excitation was used to measure emission spectra, and 600 nm emission was used to measure excitation spectra. Measurements were made in 1 PBS buffer (see section 3.4.1 for recipe) between 24–26 C. Spectra were also measured for a series of dilutions of P6 (data not shown).  3.4.3 Quantum yield measurements Fluorescein dissolved in pH 9.5 borate buffer (50 mM) with 100 mM NaCl was used as a reference with a quantum yield of 0.93.158 Fluorescein, FITC, BSA-FITC, and SCPN-P7 samples were prepared in 1 PBS buffer between 24–26 C. SCPN concentrations were from 0.25–2.5 µM. The quantum yield was determined for each material from the slopes of plots of integrated fluorescence intensity versus absorbance, relative to the slope for the fluorescein reference. The absorbance value was kept below 0.1 for the samples. The path length was 0.6 cm. The few data points for higher concentrations that appeared to deviate from linearity were excluded from calculations.  100  3.4.4 Fluorescence lifetime measurements Fluorescence lifetime measurements were made using a FluoroCube time-correlated single photon counting instrument (Horiba, Edison, NJ). Samples were measured in a 0.4 ̶ 0.7 mL (1 cm path length) quartz fluorimeter cell (Starna Cells, Atascadero, CA). The samples were prepared to a final volume of 0.4 mL by diluting in 1 PBS buffer at room temperature. The four samples for measurement were FITC (2.5 µM), fluorescein (2.5 µM), BSA-FITC (250 nM with respect to fluorescein), and SCPN-P7 (100 µM). The samples were excited with a 453 nm peak wavelength nano-LED, with a pulse duration of 1.4 ns. The fluorescence emission was selected with a 500 nm longpass filter and measured on a picosecond photon detection module (Horiba, Edison, NJ).    3.4.5 Fluorescence anisotropy measurements The fluorescence emission polarization module of the Infinite M1000 Pro plate reader was used for anisotropy measurements. Fluorescein (1.0 nM) dissolved in 0.1 M NaOH was used as the standard (27 mP) for calculating the G-factor. Excitation was at 470 nm (5 nm bandwidth) and emission measured at 520 nm (5 nm bandwidth) for fluorescein, FITC, BSA-FITC and SCPN. The same sample concentrations, buffers, and temperatures were used for spectral characterization (vide supra) and fluorescence anisotropy measurements.  3.4.6 Photobleaching measurements Photobleaching measurements were done with the same inverted microscope used for cell imaging (vide infra). Sample aliquots (5 µL) were measured in a clear flat-bottom 1536-well plate (Greiner Bio One, Kremsmünster, Austria) with the wells covered with universal optical sealing tape (#6575, Corning). Sample concentrations were adjusted to obtain similar initial fluorescence intensities (SCPN-P7, 75 µM; fluorescein, 4.1 µM; FITC, 4.4 µM; BSA-FITC, 1.0 µM with respect to fluorescein). Samples were prepared in 1 PBS buffer at room temperature. A 2 × 2 array of wells was illuminated and imaged with a 4× objective lens (NA 0.16). The excitation filter was 450/50 (center wavelength and bandwidth in nm), the dichroic mirror had a cut-off at 510 nm, and the emission filter was a longpass filter with a 500 nm cut-off (Chroma, Bellows Falls, VT). The excitation power was estimated to be ~62 mW at the sample. Images were acquired at 1 min intervals for 90 min and analyzed using ImageJ software (NIH, Bethesda, MD) with the Time Series Analyzer V3 plugin. The initial intensity for each sample was normalized to a value of unity. 101  Photobleaching rates were determined by fitting the intensity versus time data with either a monoexponential decay function (FITC, fluorescein) or a biexponential decay function (FITC-BSA, SCPN-P7).  3.4.7 Stern-Volmer quenching measurements Solutions (10 µM, 11 µL) of fluorescein, FITC, BSA-FITC, and SCPN-P7 were prepared in PBS buffer. Stock solutions (100 mM) of KI, KNO3, and KCl were prepared in 1 PBS buffer between 24–26 C. Fluorescein, FITC, BSA-FITC, and SCPN-P7 (10 μM, 11 μL) were mixed with KI, KNO3, or KCl (100 mM, 0−88 μL) and diluted to 110 μL with PBS buffer. The final concentrations of fluorescein, FITC, BSA-FITC, and SCPN-P7 were ~1.0 μM. The final concentrations of KI, KNO3, and KCl ranged between 0−80 mM. The KCl and KNO3 were used as controls to rule out any effects from ionic strength rather than the expected quenching by iodide. After 25 min incubation in the dark at room temperature, aliquots (100 μL) of these samples were transferred to a 96-well plate with 100 μL of PBS buffer as a blank. The fluorescence intensities were measured with the following excitation/emission wavelengths for each sample: 492/512 nm for fluorescein, 594/521 nm for FITC, 502/524 nm for BSA-FITC, and 492/525 nm for SCPN-P7. The excitation and emission bandwidths were 5 nm in all cases. The quenching analysis with KI was done in triplicate. Fluorescence measurements were made with the Infinite M1000 Pro multifunction plate reader and transparent, nonbinding 96-well plates (Corning, Corning, NY). The slope of the Stern- Volmer plots for KCl and KNO3 were approximately zero (data not shown).  3.4.8 Cellular viability assay The cytotoxicities of CPT-SCPN-N3 (P4) or CPT-SCPN-FA (P7) were determined with an MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt) assay kit (Abcam, Toronto, ON, Canada). The SK-BR3 cells were seeded in 96-well tissue culture-treated clear-bottom plate (ThermoFisher, Waltham, MA, USA) with a density of ~2750 cells/well and grown overnight (5% CO2, 37 °C). Cells were washed with PBS (+,+) and incubated with 1 pM ̶ 16 μM (8  10–7–1.3 mg/mL) CPT-SCPN-N3 (P4) or CPT-SCPN-FA (P7) for 2 h (5% CO2, 37 °C). After another PBS wash, the cells were replenished with fresh medium (McCoy’s) without phenol red and cultured for 3 days (5% CO2, 37 °C). After the proliferation period, 20 µL MTS reagent solution was added to each well and incubated for 2 h (5% CO2, 37 °C). 102  The absorbance was measured with an Infinite M1000 Pro plate reader (Tecan Ltd., Morrisville, NC, USA) at 490 nm and 650 nm (background). The absorbance values for each sample were background subtracted (650 nm) and the cellular viabilities were reported as a percentage of negative control wells (non-treated cells). The assays were done in triplicate.  3.4.9 Cellular immunolabeling with SCPN-Biotin (P6) SK-BR3 cells were seeded in a 96-well, tissue culture-treated, film-bottom plate (Eppendorf, Mississauga, ON, Canada). Each well was seeded with ~1.5  104 cells and the cells were allowed to grow for 2 days (5% CO2, 37 °C, McCoy’s media). Cells were washed with PBS(+, +), blocked with 1% w/v bovine serum albumin (BSA; Sigma Aldrich) in PBS(+, +) for 30 min, then incubated with 4–5 μg/mL biotinylated anti-HER2 antibody (Novus, Oakville, ON, Canada) supplemented with 1% w/v BSA for 1 h. After another wash with PBS(+, +), the cells were incubated with 0.4 ̶ 1.0 mg/mL NeutrAvidin supplemented with 0.1% w/v BSA in PBS(+, +) for 1 h, then washed with PBS(+, +) to remove excess NeutrAvidin. The cells were labeled with ~5 μM (~0.8 mg/mL) SCPN-biotin (P6) supplemented with 0.5% w/v BSA in PBS(+, +) for 1 h, and washed with PBS(+, +) prior to imaging. The entire labeling procedure was done at room temperature.  Imaging was done with an IX83 inverted epifluorescence microscope (Olympus, Richmond Hill, ON, Canada) equipped with an X-Cite 120XL metal-halide light source (Excelitas Technologies, Mississauga, ON, Canada), an Orca-Flash 4.0 V2 sCMOS camera (C11440; Hamamatsu Photonics, Hamamatsu, SZK, Japan) and motorized filter wheels (Sutter Instruments, Novato, CA, USA), and MetaMorph/MetaFluor software (Molecular Devices, Sunnyvale, CA). For cell immunolabeling, the filter set was 450/50 (center line/bandwidth in nm) for the excitation filter, a 510 nm cut-off dichroic mirror, and 520/40 or 540/50 for the emission filter. ImageJ software was used for processing images.   103   Cellular immunolabeling with a polymer dot consisting of a polyacrylate backbone polymer with pendantly functionalized blue HMAT-ODA Dye   This chapter presents work that is still being written up for publication. Refer to the Preface for full details of author contributions. Unless indicated in figure captions, I had a sole or significant role in obtaining the data presented in this chapter.   4.1 Introduction  As defined in the literature, Pdots are most often comprised of π-conjugated semiconducting polymers, which were originally developed for applications in optoelectronics including LEDs and photovoltaics.94 However, semiconducting polymers are not limited to full π-conjugation along their backbone, and may instead have active pendant groups attached along a non-conjugated backbone. The flexible backbones of the latter are attractive as they are more readily processible.205 Furthermore, their photophysical properties can be more precisely tuned to match an application.205  There are several examples that incorporate pendantly attached fluorescent dyes within polymeric nanoparticles. For example, in 2010, Achilefu et al. synthesized fluorescent crosslinked nanoparticles using a sequential one-pot functionalization/cross-linking of block copolymer micelles with amine-terminated dyes (e.g. fluorescein, Cypate) and cross-linker molecules, via reductive amination and amidation.206 In 2011, Clavier et al. utilized a polymerization-induced self-assembly (PISA) process to prepare BODIPY dye-conjugated fluorescent nanoparticles with a core/shell structure from amphiphilic copolymers.207,208 In 2013, Schubert et al. utilized nanoprecipitation to prepare fluorescent nanoparticles from poly(methyl methacrylate)-stat-poly(methacrylic acid) co-polymers with pendantly attached green, orange, or red dyes.209 These nanoparticles were prepared in the absence of surfactants (i.e. amphiphiles) and were utilized in cellular uptake studies.  104  The synthetic process demonstrated by Achilefu et al. is significantly different in comparison to that of traditional Pdots, as the dyes were covalently attached after the nanoparticles were already formed. The work demonstrated by Clavier et al. closely resembles a traditional Pdot synthesis as an emissive polymer is readily formed into a nanoparticle. However, Clavier et al. utilized an amphiphilic tri-block polymer consisting of poly(ethylene oxide), poly(acrylic acid), and a poly(styrene)-BODIPY blocks to form their particles. In contrast, traditional Pdots separate the emissive and amphiphilic polymer components, where a Pdot typically comprises >50% emissive polymer (ideally >80%). The aforementioned work demonstrated by Schubert et al. can be classified as a Pdot, as the nanoparticles are comprised of 100% emissive polymers.209 More recently, the Hudson and Algar research groups have demonstrated the development of green-emitting oxadiazole-based Pdots for applications in ratiometric oxygen sensing.210    This chapter demonstrates cellular immunolabeling with Pdots that were prepared by colleagues via coprecipitation of a blue-emitting copolymer with a polystyrene graft ethylene oxide functionalized with carboxylic acid (e.g. PS-PEG-COOH) amphiphile, as illustrated in Figure 4.1. Other examples of blue-emitting semiconducting Pdots utilized π-conjugated polymers such as polydioctylfluorene (e.g. PFO) and poly(phenylene ethynylene) (e.g. PPE). In contrast, the blue-emitting polymer utilized in this chapter consists of a hexamethylazatriangulene-oxadiazole (e.g. HMAT-ODA) emitter doped (15% w/w) within a phenyl-carbazole host on a flexible, non-conjugated polyacrylate backbone. Within the HMAT-ODA emitter, the HMAT moiety is an energy donor and ODA is an energy acceptor. The phenyl-carbazole host has a singlet energy level that is higher than the HMAT-ODA emitter, allowing for efficient energy transfer to the  HMAT-ODA. In the context of bioanalysis and bioimaging, HMAT-based emitters have shown attractive properties such as high quantum yield, good resistance to photobleaching, and amenability to two-photon excitation.211,212  For the PS-PEG-COOH amphiphilic polymer, the hydrophobic PS groups form intra- and inter-molecular bonds with other PS groups and π-conjugated groups in the HMAT-ODA polymer. The PEG chains help colloidally stabilize the Pdots and the carboxyl groups serve as distal chemical handles for bioconjugation. The HMAT-ODA Pdots are utilized for proof-of-concept extracellular labeling of the overexpressed HER2 antigen on SK-BR3 mammalian breast cancer cells. Primary 105  versus secondary immunolabeling are compared as viable approaches to specific cellular immunolabeling. These immunolabeling strategies are illustrated in Figure 4.2. HMAT-ODA Pdots were conjugated with either a mouse anti-HER2 IgG for primary immunolabeling, or with goat anti-mouse IgG for secondary immunolabeling. Well-established EDC/NHS chemistry was utilized for coupling the antibodies to the carboxyl groups present on the HMAT-ODA Pdot surface.       Figure 4.1 Preparation of antibody conjugated HMAT-ODA Pdots for cellular immunofluorescent labeling. Approximate hydrodynamic diameters: Pdots 50–70 nm. (A) Structure of blue-emitting HMAT-ODA polymer, which is an acrylate copolymer of phenylcarbazole (grey circle) and HMAT-ODA (green, and orange circles). Chris Morty Tonge synthesized the polymer. (B) Structure of amphiphilic polymer poly(styrene)-graft-poly(ethylene oxide) functionalized with carboxylic acid (PS-PEG-COOH). (C) Illustration of HMAT-ODA Pdot, Kelsi Lix synthesized the Pdots. (D) Illustration of goat anti-mouse IgG conjugated HMAT-ODA Pdot. Rupsa Gupta prepared the conjugates.   106    Figure 4.2 Strategies for immunolabeling fixed SK-BR3 cells with HMAT-ODA Pdots.  4.2 Results and discussion  4.2.1 Characterization of HMAT-ODA Pdots The Pdots were prepared via a standard nanoprecipitation technique starting from a solution of pendantly functionalized HMAT-ODA polymer and an amphiphilic polymer in THF. The amphiphile served two purposes: imparting colloidal stability and acting as a chemical handle for bioconjugation. The amphiphile used with the HMAT-ODA Pdots was PS-PEG-COOH. The mixed polymer solution was injected into water under vigorous sonication, and, following evaporation of the THF, the result was nanoparticles stabilized in water via PEG. TEM imaging of the HMAT-ODA Pdots showed spherical particles with a mean (± 1 SD) dehydrated diameter of 39 ± 39 nm (data not shown). The hydrodynamic diameter of the Pdots was measured on the 107  NTA. Two separate batches of HMAT-ODA Pdots were used to prepare primary antibody, and secondary antibody conjugated Pdots. The mean (mode) size of the corresponding batches of bare HMAT-ODA Pdots were 80 (61) ± 42 nm and 68 (52) ± 43 nm respectively.    The HMAT-ODA Pdots were covalently conjugated with either primary mouse anti-HER2 or secondary goat anti-mouse antibodies via EDC/NHS chemistry. Pdots conjugated with secondary antibody resulted in nanoparticles that had a mean (mode) hydrodynamic size of 134 (97) ± 62 nm (see Figure 4.3). We attributed the large increase in nanoparticle size in these Pdots to some degree of nanoparticle crosslinking, as expected for biomolecules linked via EDC/NHS coupling due to the large number of amine and carboxyl groups on the antibodies, and the large number of carboxyl groups on the Pdots.38    Figure 4.3 Nanoparticle tracking analysis of native HMAT-ODA/PS-PEG-COOH Pdots (black) and goat-anti-mouse IgG conjugated HMAT-ODA/PS-PEG-COOH Pdots (blue). Data contributed by Rupsa Gupta.   Photophysical properties of the HMAT-ODA Pdots are summarized in Table 4.1. Briefly, the peak wavelengths of maximum absorbance and emission are 380/50 nm (CWL/FWHM) and 450/64 nm, respectively. HMAT-ODA Pdots have a narrower emission wavelength in comparison to other Pdots such as CNMEHPPV (FWHM 134 nm), which is more ideal in multiplexing applications with multiple colors of fluorescent labels. The quantum yield of the HMAT-ODA Pdot (i.e. with amphiphile) was 0.62 ± 0.28 (± SD for batch-to-batch variation) as measured in water within an 108  integrating sphere. This quantum yield is higher than those of blue-emissive semiconducting π-conjugated Pdots such as PFO (QY 0.40) 213 and PPE (QY 0.12)213, making it ideal for applications in bioanalysis. The measured fluorescence lifetime was 4.7 ± 0.1 ns. This fluorescence lifetime was longer than those of Pdots prepared with semiconducting π-conjugated emissive polymers, which are typically in the sub-nanosecond regime.213                     Table 4.1 Fluorescence properties of HMAT-ODA Pdots.  Abs. (CWL/FWHM)a,b Em. (CWL/FWHM)a,b Φc,d Fluorescence lifetime (ns) HMAT-ODA Pdot 380/50 nm 450/64 nm 0.62 ± 0.28 4.7 ± 0.1 Notes: a CWL = Centre-wavelength, b FWHM = Full-width-half-maximum, c Quantum yield in water. d The large standard deviation is due to batch-to-batch variation. Lifetime data contributed by Ghinwa Darwish.    4.2.2 Fixed SK-BR3 immunolabeling with HMAT-ODA bioconjugates 4.2.2.1 SK-BR3 immunolabeling with primary anti-HER2 HMAT-ODA Pdot bioconjugate (direct) For immunolabeling cells with HMAT-ODA Pdots, EDC/NHS coupling was used for antibody conjugation. This chemistry targeted the lysine residues on the antibodies and the carboxyl groups of the amphiphilic PS-PEG-COOH polymer coating of the Pdots. The first immunolabeling strategy that was investigated was a direct approach using Pdots that were conjugated with primary mouse anti-HER2 antibodies as shown in Figure 4.2. This strategy was chosen first because it was simpler and does not require the use of additional antibodies for labeling. Representative results are shown in Figure 4.4. The control Pdots (i.e. non-antibody conjugated) showed a significant amount of non-specific cellular labeling, with an SBR of 2.7 ± 0.6. When cells were labeled with anti-HER2-conjugated HMAT-ODA Pdots, there was an enhanced SBR of 4.7 ± 1.5, suggesting that there was selective labeling of HER2 with conjugated antibody. The corresponding SNRs were 23 ± 5, and 27 ± 8 respectively.  109  The calculated contrast ratio (CR), comparing the SBR of the SK-BR3 cells labeled with and without HER2 antibody, was 1.8 ± 0.7. Although the signal contrast in the presence of primary antibody is promising, the high level of non-specific binding shown with the control samples is a concern and likely to be problematic with any mixed population of cells or with any multicolor labeling.     Figure 4.4 Immunolabeling of fixed, SK-BR3 cells with HMAT-ODA@anti-HER2 and HMAT-ODA (control) Pdots. Scale bars = 50 μm. A normalized pixel intensity calibration bar is provided for the relative HMAT-ODA fluorescence intensity. Images were acquired under the same microscope and camera settings. The images were adjusted to the same brightness contrast. The monochrome images were pseudo-colored cyan.   4.2.2.2 SK-BR3 immunolabeling with secondary goat anti-mouse IgG HMAT-ODA Pdot bioconjugate (indirect) A second strategy investigated for immunolabeling cells was indirect labeling with secondary antibody (Figure 4.2).94,95 The HMAT-ODA Pdots were conjugated to goat anti-mouse IgG secondary antibodies. The cells were labeled in a two-step approach, where the cells were first labeled with a primary mouse anti-HER2 antibody, then secondly with the HMAT-ODA-anti-mouse IgG Pdots. In addition, a commercial blocking agent, known as BlockAid, was included in order to help further minimize non-specific adsorption. Representative results are shown in  Figure 4.5. The SBR for SK-BR3 cells labeled in the absence of primary anti-HER2 was  0.32 ± 0.03 suggesting that there was negligible non-specific binding of HMAT-ODA-anti-mouse 110  IgG Pdots to the cells. In contrast, cells labeled in the presence of primary anti-HER2 had an SBR of 2.4 ± 0.3, which demonstrates high contrast, and specific cellular immunolabeling with the HMAT-ODA-anti-mouse IgG Pdots. Likewise, the corresponding SNRs was 3 ± 0.3, and 23 ± 2 respectively. The calculated CR, comparing the SBR of the samples with and without the primary anti-HER2 antibody was 11 ± 1, which is ~ 6-fold more robust in comparison to the CR achieved with direct immunolabeling (1.8 ± 0.7).   Figure 4.5 Differential-interference contrast (DIC, top row) and fluorescence (bottom row) microscopy images of SK-BR3 cells labeled with HMAT-ODA-Pdot-Goat anti-mouse IgG with (+) and without (─) primary labeling using anti-HER2. The images were acquired under the same microscope and camera settings. The images were adjusted to the same brightness and contrast. The monochrome images were pseudo-colored cyan.  These results suggest that the secondary antibody strategy with BlockAid better prevented non-specific cell adsorption in comparison to the direct immunolabeling method. While much of the decrease in non-specific binding for the control sample was attributed to the addition of blocking agent, there may have been some additional passivation from the conjugation of Pdots with 111  secondary antibody. Note that the control sample for the primary labeling strategy was conducted with bare HMAT-ODA Pdots, which were likely more prone to non-specific binding than antibody-conjugated Pdots. To better evaluate non-specific binding of anti-HER2 conjugated HMAT-ODA Pdots, a study that utilizes HER2-negative cells would be more suitable. Although BlockAid was not utilized in the direct immunolabeling method, the PBS buffer was supplemented with serum proteins, and magnesium and calcium chelating EDTA. The latter is commonly used for cellular suspensions in flow cytometry experiments in order to prevent non-specific binding, however, many studies use BlockAid as a blocking agent with Pdots.214–216  4.2.2.2.1 Two-photon confocal fluorescence microscopy We investigated two-photon excitation (2PE) imaging with the HMAT-ODA Pdot immunolabeled cells because it is widely used to achieve superior resolution and reduce background autofluorescence.217 To investigate the utility of the HMAT-ODA Pdots in two-photon microscopy applications, fixed SK-BR3 cells were immunolabeled via the secondary antibody approach with blocking. Representative results are shown in Figure 4.6. Interestingly, it appears that the Pdots labeled the cells both extracellularly and intracellularly. This result may be an artefact of the sample preparation as the cells were imaged through a cover slip, which had compressed the cells in the z-axis. Since the Pdots are stabilized with amphiphilic surfactant-like polymers, it is possible that these polymers act as cell membrane permeabilizing agents. The permeabilization of the membrane may introduce pores, which would allow the Pdots to enter the cells. Interestingly, cell uptake was not observed for the negative control sample, which was expected since there was little non-specific adsorption to the cells.  112   Figure 4.6 Two-photon excitation fluorescence images of SK-BR3 cells labeled with primary anti-HER2 followed by HMAT-ODA-Pdot-Goat anti-mouse IgG (left) and without primary anti-HER2 (right). The dashed white circle represents the approximate shape and location of the cell. The images represent one focal plane within in the cells. Magnification: 63×. Excitation: 770 nm (2PE). Emission: 660 nm short-pass dichroic mirror. Scale bar = 25 μm.   4.3 Conclusions   In this chapter, cellular immunolabeling with novel blue-emissive HMAT-ODA Pdots was evaluated. The HMAT-ODA Pdots exhibit good photophysical characteristics for bioimaging applications, such as relatively narrow emission spectra and high PL quantum yield. Direct and indirect antigen binding via primary and secondary antibody-conjugated Pdots were investigated. While both demonstrated specific labeling (i.e. improved contrast against control), the indirect labeling approach, aided by a blocking cocktail, demonstrated significantly lower cellular non-specific binding. Breast cancer cells immunolabeled with the HMAT-ODA Pdots were also imaged using 2PE confocal fluorescence microscopy.     113  4.4 Experimental section  The HMAT-ODA Pdots were synthesized by Kelsi Lix using nanoprecipitation procedures analogous to those we have published previously.218 The HMAT-ODA Pdot antibody bioconjugates were prepared by Rupsa Gupta, using carbodiimide coupling methods similar to those published previously.96  4.4.1 Immunolabeling of fixed SK-BR3 cells with fluorescent nanoparticles 4.4.1.1 Preparation of fixed SK-BR3 cells 4.4.1.1.1 Cell culture SK-BR3 cells were cultured in a humidified incubator with 95% air/5% CO2 at 37 °C. The culture medium was McCoy’s 5A (GE Healthcare, Chicago, IL) supplemented with 10% v/v fetal bovine serum and 1× antibiotic and antimycotic (ThermoFisher). Cells were cultured in T25 flasks and sub-cultured every 5–7 days.   4.4.1.1.2 Cell fixation A suspension of freshly trypsinized SK-BR3 cells (1.4 × 106 cells) was pelleted by centrifugation at 55 RCF for 5 min. The supernatant was removed, and the pellet resuspended in 2 mL of  1 × PBS buffer (Gibco, pH 7.2, 1.54 mM KH2PO4, 2.71 mM Na2HPO4, 155 mM NaCl). A volume of 2 mL of 4% (w/v) paraformaldehyde in PBS was added and the sample gently mixed via pipette. The sample was incubated at room temperature for 5–10 min before pelleting via centrifugation at 55 RCF for 5 min. The supernatant was discarded, and the pellet resuspended in 2 mL of 1 × PBS buffer. The cells were then pelleted and resuspended in 2 mL of labeling buffer (1% BSA, 1 mM EDTA in 1 × PBS buffer).  4.4.1.2 HMAT-ODA Pdots 4.4.1.2.1 Cell immunolabeling direct approach Samples were prepared by incubating 0.1 mL of 5 × 105 cells/mL stock with ~10 pM HMAT-ODA-Pdot sample (+ anti-HER2, without IgG), which were both diluted in supplemented PBS buffer (2% v/v FBS, 1 mM EDTA 1× PBS) prior to introducing the Pdots to the cells. The cells were incubated with the Pdots in the dark on the rotary mixer for 10 minutes. After mixing, the 114  cells were pelleted at 80 rcf for 4 minutes, supernatant was removed, and the pellet was washed in 0.3 mL supplemented PBS buffer. The sample was then centrifuged again, and supernatant was removed until ~10 µL remained. Then 10 µL supplemented PBS was added, and the sample was mixed. For microscope imaging, a 7.5 µL droplet of cell suspension was applied to a glass slide, and a cover slip was applied.  4.4.1.2.2 Cell immunolabeling indirect approach A 50 µL suspension of fixed SK-BR3 cells (7 × 105 cells/mL) in labeling buffer (1% BSA, 1 mM EDTA in 1 × PBS buffer) was incubated with 5 µg/mL primary antibody (anti-HER2, Novus Biologicals) on a rotary shaker for 30 min in the dark at room temperature. The labeled cell suspension was then pelleted briefly via centrifugation, and the pellet was washed with 2 x 100 µL aliquots of labeling buffer. Then the labeled cells were incubated with 70 pM Pdot-IgG conjugates (0.5 µg/mL goat anti-mouse IgG) in BlockAid buffer for 15 min on the shaker in the dark at room temperature. A control sample was prepared with unlabeled cells. Following incubation, the cells were then pelleted briefly via centrifugation, and the pellet was washed with 2 x 100 µL labeling buffer. The cells were resuspended in 20 µL of labeling buffer, and a drop of cell suspension was placed on a microscope slide, a cover slip was applied, and the cells were imaged through the cover slip under a fluorescence microscope (Olympus IX83).  4.4.1.2.3 Fluorescence microscopy Cell imaging was conducted on a fluorescence microscope. This microscope was an IX83 inverted microscope (Olympus, Richmond Hill, ON, Canada) equipped with an X-Cite 120XL metal-halide light source (Excelitas Technologies, Mississauga, ON, Canada), a white-LED transmitted light source, an Orca-Flash 4.0 V2 sCMOS camera (C11440; Hamamatsu Photonics, Hamamatsu, SZK, Japan), motorized filter wheels (Sutter Instruments, Novato, CA), and MetaMorph/MetaFluor software (Molecular Devices, Sunnyvale, CA). For cell immunolabeling, the filter set was 350/50 for the excitation filter (center line/bandwidth in nm), a 425 nm cut-off dichroic mirror, and a 460/50 bandpass emission filter. Filters and dichroic mirrors were from Chroma (Bellows Falls, VT). Images were acquired using a 10× objective lens (UPLSAPO10X2, NA: 0.4, Olympus) or a 60× objective lens (UPLFLN60X, NA: 0.9, Olympus). Images were acquired with an exposure time of 50 ms. 115  4.4.1.2.4 Two-photon microscopy The samples above were also imaged using two-photon confocal fluorescence microscopy. This microscope was a Zeiss LSM 510 MP. The tunable femtosecond laser used for sample illumination was Chameleon Ultra femtosecond laser (Coherent Inc.). The images were acquired at a magnification of 63X under oil-immersion (Zeiss, Plan-Neofluar, NA = 1.40). The sample was excited at 770 nm. A 660 nm shortpass filter was used to block out the excitation source and select for the emitted blue light from the HMAT-ODA Pdots. Images were acquired at a z-depth of 1 µm within the cells.  116   Cellular immunolabeling with semiconducting π-conjugated polymer dots   This chapter is an adaptation Lix, K.; Tran, M.V.; Massey, M.; Rees, K.; Sauvé, E.R.; Hudson, Z.M.; Algar, W.R., Dextran Functionalization of Semiconducting Polymer Dots and Conjugation with Tetrameric Antibody Complexes for Bioanalysis and Imaging. ACS Appl. Bio Mater. 2020, 3, 432–440, with permission from the American Chemical Society (Copyright 2020). Refer to the Preface for full details of author contributions. Unless indicated in figure captions, I had a sole or significant role in obtaining the data presented in this chapter.   5.1 Introduction  In contrast to Chapter 4, this chapter will explore using Pdots composed of π-conjugated semiconducting polymers for applications in immunofluorescent labeling. Pdots are typically held together by intra- and intermolecular hydrophobic interactions. Colloidal stabilization of Pdots is typically via amphiphilic polymers, which have hydrophobic residues that interact with the hydrophobic semiconducting polymers and hydrophilic groups that provide colloidal stabilization via electrostatic and/or steric repulsion. Furthermore, the hydrophilic groups provide a chemical handle, such as carboxylic acid groups, for the bioconjugation of biomolecules. In the case of amphiphiles with only carboxylic acid groups, colloidal stability is typically poor in solutions with low pH or high ionic strength. Furthermore, Pdots also have a propensity to non-specifically bind to some surfaces, proteins, and cells. The development of Pdot surface chemistry that imparts robust colloidal stability and low non-specific binding is essential for taking full advantage of fluorescence properties of these materials.   In biological applications requiring specific targeting, antibodies specific to extracellular antigens are commonly conjugated to Pdots. This bioconjugation frequently occurs through EDC/NHS chemistry, which covalently links a carboxylic acid on the Pdot surface with lysine residues of the antibody. Limitations of this chemistry include poor control of antibody orientation (i.e. the epitope-binding region may be occluded by the nanoparticle), being prone to interparticle 117  crosslinking, and poor reproducibility. Improved methods for the well-controlled conjugation of antibodies to Pdots are needed.  This chapter demonstrates cellular immunolabeling using a combination of dextran-coated Pdots and TACs. (The advantages of dextran as a coating material were introduced in Chapter 2) Pdots were prepared by nanoprecipitation with a PSMA amphiphilic polymer. The resulting surface carboxylic acid groups were conjugated and partially crosslinked with an amine-modified dextran. Both green-emitting poly(9,9-dioctylfluorene-alt-benzothiadiazole) (F8BT) or red-orange-emitting poly[2-methoxy-5-(2-ethylhexyloxy)-1,4-(1-cyanovinylene-1,4-phenylene)] (CNMEHPPV) semiconducting polymers were used to prepare the Pdots. A schematic illustrating the full preparatory route is shown in Figure 5.1. The dextran coating imparted superior colloidal stability to the Pdots in comparison to the amphiphilic polymer alone and acted as a biomolecular anchor for TACs. TACs are commonly used with magnetic nanoparticles for cellular isolation applications;75,219 however, there has been little investigation into their usefulness with fluorescent probes. These TACs utilize dextran-specific and HER2 specific antibodies and are advantageous as they direct the cell targeting antibody (i.e. HER2) in an orientation that is favorable for cellular binding. This bioconjugation method is versatile as the TACs can potentially be prepared using different cell-targeting antibodies, allowing for quick plug-and-play capabilities with dextran-functionalized semiconducting Pdots. The TAC-based bioconjugation strategy is illustrated in Figure 5.2.    118    Figure 5.1 Preparation of dextran-functionalized F8BT or CNMEHPPV Pdots for immunofluorescent cellular labeling. Approximate hydrodynamic diameters: Pdots 50–70 nm. Pdots were synthesized by Kelsi Lix.    119    Figure 5.2 Strategies for immunolabeling fixed SK-BR3 cells with dextran-coated F8BT and CNMEHPPV Pdots.  5.2 Results and discussion  5.2.1 Characterization of F8BT or CNMEHPPV Pdots Pdots were prepared via a standard nanoprecipitation method. Semiconducting polymer (e.g. F8BT, CNMEHPPV) and an amphiphilic polymer were dissolved in THF and this solution was injected into water under vigorous sonication, resulting in nanoparticles after evaporation of the THF.95 The amphiphile serves two purposes: imparting colloidal stability via electrostatic repulsion or sterics, and acting as a chemical handle for functionalization with aminated dextran. The amphiphile was polystyrene-co-maleic anhydride (PSMA), which initially stabilized the nanoparticles in water electrostatically as the maleic anhydride groups hydrolyzed to carboxylic acids. The mean (mode) hydrodynamic sizes of the F8BT and CNMEHPPV Pdots were 51 (39) ± 23 nm and 54 (39) ± 1 nm, respectively, as determined by NTA.  120  The immunoconjugation of the F8BT and CNMEHPPV Pdots via TACs required their functionalization with TAC-supporting dextran coating, which also improved nanoparticle stability. The dextran (6 kDa) was first modified with amine groups via reductive amination with hexamethylenediamine. The F8BT and CNMEHPPV Pdots were then conjugated with the amine-functionalized dextran via EDC/NHS coupling at their carboxylic acid groups. The mean (mode) hydrodynamic diameter of the dextran-coated Pdots (Dex-Pdots) prepared with F8BT was  59 (44) ± 28 nm when measured by NTA. Prior to immunolabeling studies, the Dex-Pdots were conjugated to TAC-HER2 at a molar ratio of 21:1 TAC-HER2:Dex-Pdot.  Photophysical properties of the Pdots are listed in Table 5.1. Both Pdots had a similar absorbance maximum with relatively broad FWHMs. Of the two materials, the F8BT Pdots had a significantly narrower emission band and green-yellow emission, whereas the CNMEHPPV Pdots had a broader emission band with orange-red emission. The two materials had similar quantum yields, albeit with large batch-to-batch variations.                           Table 5.1 Fluorescence properties of F8BT and CNMEHPPV Pdots. Pdot Abs.  (CWL/FWHM) a,b Em. (CWL/FWHM)a,b Φc,d F8BT 468/94 nm 540/78 nm 0.35 ± 0.24 CNMEHPPV   468/100 nm 624/134 nm 0.37 ± 0.30 Notes: a CWL = Centre-wavelength, b FWHM = Full-width-half-maximum, c Quantum yield. d The large deviation is due to batch-to-batch variation. Quantum yield data was contributed by Kelsi Lix.   5.2.2 Fixed SK-BR3 immunolabeling with F8BT/CNMEHPPV Pdot bioconjugates via TAC-HER2 TAC conjugates of Dex-Pdots (F8BT or CNMEHPPV) were used to immunolabel fixed SK-BR3 cells, which overexpress the HER2 antigen. The TAC incorporated anti-dextran and anti-HER2 antibodies, as illustrated in Figure 5.2. There was an average of 21 TACs per Pdot. Representative images of cells are shown in Figure 5.3 for F8BT Dex-Pdots. Negligible non-specific binding to the cells in the absence of TAC-HER2 was demonstrated, with calculated signal-to-background 121  (SBR) ratios of ~ 0. In contrast, cells that were labeled with the TAC-HER2-conjugated Dex-Pdots showed an SBR of 1.0 ± 0.1 for the F8BT Dex-Pdots. The corresponding SNR ratio was 6.5 ± 0.6.     Figure 5.3 Immunolabeling of fixed, and DAPI (nuclear stain) labeled SK-BR3 cells with F8BT/Dex Pdots via TAC-HER2 (bottom row). Scale bars = 50 μm. A normalized pixel intensity calibration bar (below) is provided for the relative F8BT fluorescence intensity. Images were acquired under the same microscope and camera settings. The images were adjusted to the same brightness contrast for the different color channels. The monochrome images were pseudo-colored green (F8BT) or blue (DAPI).    In a separate experiment, the cell immunolabeling efficiency of Dex-F8BT Pdots was compared to unmodified F8BT Pdots (see Figure 5.4). Interestingly, there was significantly more cell labeling for unmodified F8BT Pdots (no dextran) in the absence of TAC-HER2, suggesting that the functionalization with dextran helps prevent non-specific interactions with cells. Furthermore, the significant increase in TAC-mediated cellular labeling with Dex-Pdots (SBR 8.1 ± 2.3, SNR 122  30 ± 8) relative to unmodified Pdots (SBR 1.7 ± 0.6, SNR 8 ± 3) indicates that the dextran modification is required for efficient labeling.     Figure 5.4 Comparison of fixed SK-BR3 immunolabeling with dextran-functionalized (F8BT/Dex) and un-modified (F8BT) Pdots via TAC-HER2. A normalized pixel intensity calibration bar (below) is provided for the relative F8BT fluorescence intensity. Images were acquired under the same microscope and camera settings. The images were adjusted to the same brightness contrast for the different color channels. The monochrome images were pseudo-colored green (F8BT).   Dextran-functionalized CNMEHPPV Pdots showed similar cellular immunolabeling results in comparison to their F8BT counterparts. As shown in Figure 5.5, the calculated SBR in the presence of TAC-HER2 was 1.5 ± 0.1, whereas the control sample showed negligible non-specific cellular binding, and a calculated SBR of ~ 0. The corresponding SNR was 13 ± 1 for the TAC-HER2 positive sample. These results suggest that the dextran coating helps prevent the Pdots from adsorbing to the cells, and also provides a biochemical handle for TAC-mediated cellular immunolabeling. The specificity of this TAC-HER2 complex is further studied in Chapter 7, whereby cellular labeling and isolation is compared for both HER2 positive and negative cells. 123     Figure 5.5 Immunolabeling of fixed, and DAPI (nuclear stain) labeled SK-BR3 cells with CNMEHPPV//Dex Pdots via TAC-HER2 (bottom row). Scale bars = 50 μm. A normalized pixel intensity calibration bar (below) is provided for the relative CNMEHPPV fluorescence intensity. Images were acquired under the same microscope and camera settings. The images were adjusted to the same brightness contrast for the different color channels. The monochrome images were pseudo-colored red (CNMEHPPV) or blue (DAPI).  5.3 Conclusion  This chapter has presented a method for immunolabeling breast cancer cells using dextran-functionalized Pdots and TACs. The dextran coating significantly reduced non-specific binding associated with the Pdot and enabled immunoconjugation via TACs, the utility of which was demonstrated through labeling of HER2-positive SK-BR3 cells. Dextran functionalization is a promising strategy for overcoming some the current limitations of Pdots, such as modest stability, tendency toward non-specific binding, and limited bioconjugate chemistries.  124  5.4 Experimental section  5.4.1 Synthesis of Pdots 5.4.1.1 F8BT and CNMEHPPV The methods for the synthesis and preparation of dextran-functionalized F8BT and CNMEHPPV Pdots can be found in the published paper.218  5.4.2 Immunolabeling of fixed SK-BR3 cells with F8BT/CNMEHPPV Dex-Pdots Fixed SK-BR3 cells were prepared using the method described earlier in section 4.4.1.1  5.4.2.1.1 Preparation of (Anti-HER2-TAC)-Dex-Pdot Conjugates A 15 µg/mL stock solution of TAC with anti-HER2 antibody was prepared in PBS buffer by following the manufacturer’s protocol for the Do-It-Yourself Positive Selection Kit II. Briefly, 15 µg of mouse anti-HER2 antibody (NBP2-32863; Novus Biologicals, Centennial, CO, USA) was mixed with 100 µL of Component A and 100 µL of Component B (from kit). The sample was incubated overnight at 37 ˚C. The TAC-anti-HER2 complex was then diluted to 1.0 mL with PBS buffer and stored at 4 ˚C until needed.   TAC-Pdot conjugates were freshly prepared prior to cell immunolabeling. The TAC-anti-HER2 complex was diluted to a final concentration of 22 nM in 1× PBS buffer. An aliquot of TAC stock (30 µL, 22 nM, 0.646 pmol, 21 equivalents) was spiked into a 100 µL solution of Dex-Pdots (0.03 pmol, 300 pM). The sample was topped up by the addition of 20 µL of 1× PBS buffer, and then incubated at room temperature for 30 min.    5.4.2.1.2 SK-BR3 immunolabeling and cell imaging A suspension of freshly trypsinized SK-BR3 cells (~106 cells) was pelleted by centrifugation at  55 RCF for 5 min. The supernatant was removed, and the pellet resuspended in 2.0 mL of PBS buffer. A volume of 2.0 mL of 4% (w/v) paraformaldehyde in PBS was added and the sample gently mixed via pipette. The sample was incubated at room temperature for 5–10 min before pelleting via centrifugation at 55 RCF for 5 min. The supernatant was discarded, and the pellet resuspended in 4.0 mL of PBS buffer.  125  A suspension (20 µL) of paraformaldehyde-fixed SK-BR3 cells in PBS was pipetted into a 1.7 mL microcentrifuge tube. The SK-BR3 cell suspension was then spiked with 5.0 µL of (anti-HER2-TAC)-Dex-Pdot conjugates (~300 pM, 21 equiv. TAC-anti-HER2 relative to nanoparticles). The sample mixture was mixed briefly via pipette and then incubated on the benchtop for 15 min in the dark. Following incubation, the labeled cells were pelleted via centrifugation at 55 RCF and the supernatant removed by pipette. The cells were resuspended in 20 µL of fresh PBS buffer.   Samples were prepared by pipetting 7.5 µL of a suspension of labeled cells onto a microscope slide. A cover slip was applied, and the sample was inverted and imaged through the cover slip. The fluorescence filter sets used for imaging were as listed in Table 5.2. The emission spectra of the cells labeled with (anti-HER2-TAC)-Dex-Pdot conjugates (F8BT) were acquired with a diode-array spectrometer (Greenwave 16 VIS-50; StellarNet, Tampa, FL) that was coupled to the trinocular head of the microscope via a fiber-optic cable.   Table 5.2 Fluorescence microscopy optics for SK-BR3 labeling. (anti-HER2-TAC)-Dex-Pdot Ex. Filter a Em. Filter a, b Dichroic Mirror c Objective Lens d F8BT 450/50 BP 500LP T470 100XO (1.40 NA) CNMEHPPV 450/50 BP 550LP T565 60X (0.9 NA) Notes: a Center wavelength/bandwidth, BP = bandpass filter. b LP = longpass filter. c T = transmission cut-on wavelength. All numbers in units of nanometers. d X = magnification factor, air-immersion; XO = magnification factor, oil-immersion; NA = numerical aperture.  126   Fully self-assembled silica nanoparticle-semiconductor quantum dot supra-nanoparticles and immunoconjugates for enhanced cellular imaging by microscopy and smartphone camera   This chapter is an adaptation Darwish, G.H.; Asselin, J.; Tran, M.V.; Gupta, R.; Kim, H.; Boudreau, D.; Algar, W.R., Fully Self-Assembled Silica Nanoparticle-Semiconductor Quantum Dot Supra-Nanoparticles and Immunoconjugates for Enhanced Cellular Imaging by Microscopy and Smartphone Camera. ACS Appl. Mater. Interfaces 2020, 12, 33530–33540, with permission from the American Chemical Society (Copyright 2020). Refer to the Preface for full details of author contributions. Unless indicated in figure captions, I had a sole or significant role in obtaining the data presented in this chapter.   6.1 Introduction  Many luminescent materials are available for applications in bioanalysis and imaging, including conventional organic fluorescent dyes and fluorescent proteins,220,221 conjugated polymer nanoparticles,222 semiconductor QDs,39 upconverting and downshifting lanthanide-based nanoparticles, metal nanoclusters, 223 various carbon nanomaterials,224–226 and several others. Each material exhibits intrinsic photophysical properties, which may be beneficial in some applications but detrimental in others. These properties include brightness, emission rate and bandwidth, spectral regions for excitation and emission, blinking, resistance to photobleaching, robustness, and other properties. In scenarios where a specific label displays photophysical properties that are advantageous for an application, but greater sensitivity is still desired, a strategy has been the preparation of brighter materials through incorporation of multiple copies of a fluorescent label in a single-particle vector. Early examples of this strategy include silica and polymer NPs that were doped with molecular dyes.227,228 More recent examples have extended this concept by doping larger luminescent NPs within host NPs, or on their surface, as is the case for supra- and super- nanoparticle (NP) assemblies with luminescent NPs.229  127  Brighter materials provide greater sensitivity, thus benefiting applications that have signals restricted due to the low abundance of a target biomarker, intrinsic background from the sample (e.g. autofluorescence background, light scattering), and/or utilize detection devices that provide greater portability at a lower cost but at the expense of sensitivity. Examples of the latter are POC diagnostic tests that utilize luminescence detection, often prioritizing rapid results, low cost, and simple on-site testing.230–232 POC diagnostic technologies have recently leveraged smartphones as a powerful tool for biomarker detection and/or interfacing other devices.233–235 Possibly the greatest challenge with adapting smartphones for POC diagnostic tests is that their cameras are neither designed nor optimized for fluorescence imaging. These challenges are being overcome with sophisticated engineering solutions, and as important, material solutions that leverage the photophysical properties of different luminescent NPs.111 There is also an increasing need for improved detection of low-abundance biomarkers within the research laboratory, including contexts such as cellular imaging, flow cytometry, in vitro assays, and other methods.  Colloidal semiconductor QDs are luminescent NPs that continue to be of interest for bioanalysis and imaging.36,236,237 QDs offer significant improvements in photostability in comparison to both organic fluorescent dyes and Pdots.64 Another advantage of QDs is the added capability of multiplexing, as their broad absorption bandwidth coupled with their narrow emission bandwidths allows multiple colors of QDs to be excited and resolved simultaneously. This multiplexing capability reduces the complexity of the optical equipment required for cellular imaging, making them suitable for prospective POC diagnostic technologies. Despite these advantages, single QDs may lack sufficient sensitivity when utilized in applications where biomarkers are in low-abundance, or readout technology is non-optimal, such as in POC diagnostic technologies. However, nanoparticle architectures that constitute many fluorescent reporters in a single-particle vector may mitigate this challenge by increasing the relative particle brightness for improved sensitivity in bioanalytical applications. Examples of this include composites of QDs in the form of polymer NPs doped with QDs238 and, more recently, supra-NP and super-NP assemblies with QDs. Supra-NP assemblies provide a good balance between simple preparation, particle size, and total number of QDs.239,240 Supra-NPs require a scaffold particle for assembling QDs. Silica NPs are potentially ideal for this purpose because of their water compatibility, robustness, optical transparency, and amenability to a suite of chemical modifications.17,241,242 Examples of silica NP-128  based supra-NP assemblies include applications such as cellular and in vivo imaging;240 and immunoassays for peptides, proteins243–245 and pathogens.246  This chapter demonstrates cellular immunolabeling with supra-NP assemblies, SiO2@QDs, that comprise a silica NP coated with a corona of dextran-functionalized QDs, as shown in Figure 6.1. The SiO2@QDs were prepared fully by self-assembly: QDs bound via electrostatic and/or coordinate interactions to silica NPs that were chemically modified with imidazoline moieties; dextran was chemically modified with imidazole groups (API) that bound to the QDs through coordinate interactions; and immunoconjugation was via a bispecific TAC with anti-dextran and anti-cell antibodies. The cellular immunolabeling strategy is illustrated in Figure 6.2.     Figure 6.1 Scheme for the preparation of self-assembled SiO2@QD supra-nanoparticles and their functionalization with dextran. SiO2 NPs are modified with triethoxy-3-(2-imidazolin-1-yl)propylsilane (IPS), where the imidazoline groups spontaneously bind (electrostatically or coordinate) CdSe/CdS/ZnS QDs coated with glutathione ligands (not shown). The QDs are then further functionalized with dextran that has been modified with pendant 1-(3-aminopropyl)imidazole groups. The SiO2 NPs and QDs are drawn approximately to scale. The IPS and dextran molecules are not drawn to scale. The SiO2@QD were prepared by Ghinwa Darwish and Jérémie Asselin.  129    Figure 6.2 Strategies for immunolabeling fixed SK-BR3 cells with dextran-coated SiO2@QDs.   6.2 Results and discussion  6.2.1 SK-BR3 immunolabeling with SiO2@(QD-Dex) supra-nanoparticles 6.2.1.1 Preparation of SiO2@(QD-Dex) supra-nanoparticles Figure 6.1 illustrates the preparatory route for the colloidal SiO2@QD supra-NP self-assemblies. Silica NPs were first synthesized using a modified Stöber method,247 then modified with an imidazoline-terminated silane. The imidazoline group was selected because of its similarity to imidazole groups, which are well-known to bind to the inorganic surface of QDs, most frequently in the form of polyhistidine tags or polymers with multiple pendant imidazole groups. When the imidazoline-functionalized silica NPs were mixed with glutathione (GSH) ligand-functionalized QDs, a monolayer of QDs assembled at the surface of the silica NPs. The QDs could then be 130  further assembled with 1-(3-aminopropyl)imidazole-modified dextran (API-Dex), which was useful for enhancing colloidal stability and for subsequent immunoconjugation with TACs (vide infra).   6.2.1.2 Characterization of SiO2@QD supra-nanoparticles 6.2.1.2.1 Physical characterization The SiO2@QD supra-NP assemblies were characterized by TEM, NTA, fluorescence spectroscopy, and zeta-potential measurements.   The size distribution of the precursor SiO2 NPs as determined by TEM was 76 ± 9 nm (± one standard deviation). The hydrodynamic diameter, as measured by scattering mode NTA, was  83 ± 21 nm for the SiO2 NPs. SiO2@QD625 supra-NPs with maximum surface coverage of GSH-coated CdSe/CdS/ZnS QD625 (i.e. QDs with a peak PL at ~625 nm) were characterized by TEM. The TEM images in Figure 6.3 show the QDs as smaller and darker spherical particles distributed on the surface of the larger and less dark SiO2 NPs. The average number of QDs per SiO2 NPs was extracted from the TEM images, yielding a value of 37 ± 8 QDs per SiO2 NP. Since the TEM images only display the QDs on one side of the particles, the number of particles counted was multiplied by two. This assumption was made because there were relatively few instances of overlap of QDs within each of the supra-nanoparticles. Given the density of QDs, many instances of overlap would have been expected if both faces of the particles were clearly observed in the TEM images.  The hydrodynamic size distribution of fluorescent SiO2@QD625 supra-NPs was characterized using both the scattering, and fluorescent modes of the NTA (data not shown). The reported mean hydrodynamic diameters of the particles were 111 ± 36 nm and 110 ± 38 nm by scattering and fluorescence measurement modes, respectively, which, as expected, was larger than the hydrodynamic diameter of the SiO2 NPs (83 ± 21 nm). 131    Figure 6.3 TEM image of SiO2@QD supra-nanoparticles. The inset is a higher magnification image. Scale bar represents 100 nm. Inset scale bar represents 50 nm. TEM images contributed by Ghinwa Darwish and Jérémie Asselin.   The assembly of QDs on SiO2 NPs was confirmed by zeta-potential measurements for the nanoparticles at various stages of the supra-NP preparation. The initial, as-synthesized SiO2 NPs had a zeta-potential of –11 mV from ionized silanol groups. After the functionalization of the  SiO2 NPs with imidazoline groups, the zeta-potential increased to + 13 mV, which is consistent with the protonation of the imidazoline (pKa: ~11) at a sample pH of 8.5.248 Upon assembly of the GSH-coated QDs, the zeta-potential decreased to –29 mV for a sample with a maximum loading of QDs. The negative zeta-potential resulted from the anionic character of the QDs, which arose from the carboxylate groups of their GSH ligands.  6.2.1.2.2 Photophysical characterization The absorbance, emission, and excitation spectra were measured for ensembles of both single GSH-QD625 and SiO2@QD625 supra-NPs. The absorbance spectra of the supra-NPs resembled the light scattering spectrum for SiO2 NPs. The PL excitation spectra, which ensures that the SiO2 132  NP scattering effects are negligible, demonstrated the first and second excitonic peaks in the supra-NPs centered at ~625 nm and ~577 nm respectively. The emission spectrum for the supra-NPs showed a modest 2 nm bathochromic spectral shift and 10% improvement in apparent PL intensity relative to single QDs. Fluorescence lifetime measurements showed average PL lifetimes of were 27.5 ± 0.4 ns for QD625 and 21.1 ± 2.3 ns for SiO2@QD625, which is consistent with the relative PL intensity enhancement for the supra-nanoparticles relative to single QDs. When ensembles of single QDs and supra-NPs were illuminated continuously, both materials underwent photobrightening within a 60-min period but did not photobleach. In single-particle PL imaging experiments, SiO2@QD625 were ~13-fold brighter than single GSH-QD625s. Intensity-time trajectories for single-particle imaging of these particles revealed blinking (i.e. PL intermittency) for single QDs but no blinking for SiO2@QD625s.  6.2.1.3 Immunolabeling with SiO2@(QD-Dex) supra-nanoparticles 6.2.1.3.1 Microscope-based cellular imaging To demonstrate their potential utility as fluorescent labels for cellular imaging, batches of SiO2@QD605 and QD605 were used to immunolabel fixed mammalian breast cancer cells  (SK-BR3) that overexpress the HER2 on their surface. For this application, both the SiO2@QD605 and QD605 were coated with API-modified dextran (API-Dex, 10 kDa) using methods similar to those we have reported in Chapters 2 and 7,249,250 yielding Dex-QD605 and SiO2@(QD605-Dex). SiO2@(QD605-GSH) (no dextran) was also tested as a control. Immunolabeling was done via a TAC with both anti-dextran and anti-HER2 antibodies. As illustrated in Figure 6.2, this TAC spontaneously and selectively binds dextran-functionalized QDs or SiO2@QDs to cells expressing HER2. As controls, cells were exposed to all three materials, both with and without added TAC.   Figure 6.4A shows brightfield and PL images of cells that were nominally immunolabeled with Dex-QD605, SiO2@(QD605-GSH), and SiO2@(QD605-Dex). Images were acquired and processed with the same settings. DAPI was used as blue-fluorescent stain for the cell nuclei. Immunolabeling with TAC+(Dex-QD605) resulted in very low but measurable QD PL intensities, and there was no detectable QD PL intensity without TAC when images were acquired under the same illumination conditions and camera settings. Much higher QD PL intensity was observed for nominal immunolabeling with SiO2@(QD605-GSH)+TAC; however, high QD PL intensity was 133  also observed without TAC, indicating significant non-specific binding to cells. Immunolabeling with SiO2@(QD605-Dex)+TAC yielded the highest QD PL intensity, as well as low PL intensity without TAC. Figure 6.4B shows that the contrast ratio between cell samples with and without TAC was a meager 1.9:1 for SiO2@(QD605-GSH), but a robust 6.6:1 with SiO2@(QD605-Dex). Figure 6.4C shows lower magnification images of cells that were TAC immunolabeled with these materials, where the signal-to-background ratio (SBR) was 1.7 ± 0.7 (± 1 SD) for SiO2@(QD605-GSH) versus 7.7 ± 5.3 for SiO2@(QD605-Dex). The corresponding SNR values were 12 ± 5 and 51 ± 35.   Figure 6.4 (A) TAC-based immunolabeling of HER2-expressing SK-BR3 cells with Dex-QD605, SiO2@(QD605-GSH), and SiO2@(QD605-Dex), and control experiments without TAC. Brightfield images are shown alongside DAPI (nuclear stain) fluorescence images and QD PL images. A pixel-intensity calibration bar is shown. Scale bar = 20 µm in all images. (B) Comparison of the contrast ratios for specific labeling (with TAC) and non-specific binding (without TAC) for SiO2@(QD605-GSH) and SiO2@(QD605-Dex). For comparison, 10 cells were analyzed for each sample. (C) Images highlighting the difference in signal-to-background ratio between specific labeling (with TAC) with SiO2@(QD605-GSH) and SiO2@(QD605-Dex). The scale bars = 100 µm and a pixel-intensity calibration bar is shown. 134  6.2.1.3.2 Smartphone-based cellular imaging To further evaluate the cellular imaging potential of the SiO2@QD supra-NPs, an application that would specially leverage their high per-particle brightness was sought. Smartphone-based fluorescence assays and imaging are emerging as a promising platform for point-of-care molecular and cellular diagnostics and are an application that needs high-brightness materials because the smartphone camera is neither designed nor optimized for fluorescence imaging. Fixed SK-BR3 breast cancer cells were again immunolabeled using Dex-QD605, SiO2@(QD605-GSH), and SiO2@(QD605-Dex), both with and without TAC. The cells were then imaged on the platform shown in Figure 6.5A, the design of which will be discussed in Chapter 7.250 Figure 6.5B shows the resulting PL images. Many individual cells were barely detectable when labeled with TAC+(Dex-QD605), with an SBR of 0.4 ± 0.1. Only a few cells were detectable without TAC. Cells labeled with SiO2@(QD605-GSH)+TAC were much brighter, with an SBR of 1.4 ± 0.4; however, there was again significant non-specific binding without TAC, as indicated by the large number of detected cells. Labeling with SiO2@(QD605-Dex)+TAC provided highest signal-to-background ratio, 3.1 ± 1.3, and there was minimal non-specific binding in the absence of TAC. The corresponding SNRs were 3 ± 0.5, 6 ± 2, and 9 ± 4.      Figure 6.5 (A) Cross-sectional view for the design of smartphone-based platform for PL imaging of cells. Details of the design can be found in Chapter 7. (B) Smartphone-acquired images of SK-BR3 cells that were TAC-immunolabeled with Dex-QD605, SiO2@(QD605-GSH), and SiO2@(QD605-Dex), and control experiments without TAC. Scale bars = 500 μm.   135  6.2.1.3.3 Immunolabeling with multiple colors of SiO2@QD The next capability investigated was the potential for multiplexed/multicolor imaging utilizing SiO2@QD supra-NPs. The self-assembly process made it easy to prepare multiple colors blue, green, and red supra-NPs, each coated with API-Dex. For proof of concept with each color, a single antigen (e.g. HER2) on SK-BR3 cells was immunolabeled and imaged on the smartphone imaging platform in Figure 6.6. The results demonstrate excellent immunolabeling contrast between cells mixed with SiO2@(QD-Dex) in the presence and absence of TAC-HER2. In the presence of TAC-HER2, the signal-to-background ratio (mean ± 1 SD) was 2.2 ± 1.1, 0.7 ± 0.2, and 0.7 ± 0.2 for cells labeled with blue, green, and red supra-NPs, respectively. The corresponding SNRs were 9.6 ± 4.7, 4.8 ± 0.1, and 4.3 ± 1.0 respectively. In the absence of TAC-HER2, there was a negligible amount of labeling for all colors of supra-NPs, with an SBR of ~ 0. Furthermore, the number of cells counted in the presence of TAC-HER2 was significantly greater in comparison to the negative control. With TAC-HER2, the total number of cells counted was 340, 248, and 242 for cells labeled with blue, green, and red supra-NPs, respectively, corresponding to 16-fold, 19-fold, and 19-fold improvements in the number of cells counted. The cells counted in the control samples without TAC-HER2 were either anomalous non-specifically labeled cells and/or aggregates of the supra-NPs. Once a method using different TACs in a parallel is developed, the red, green, and blue SiO2@QD will be useful for profiling multiple cell types within a heterogeneous population of cells on a smartphone imaging platform.    136    Figure 6.6 Smartphone-based imaging of fixed SK-BR3 cells immunolabeled with SiO2@(QD-Dex) supra-nanoparticles without (─) TAC-HER2, and with (+) TAC-HER2.The QD colors are denoted as QD λ, where λ represents the center peak wavelength of the QD. Scale bars = 500 μm.   6.3 Conclusion   In this chapter, a facile method for the self-assembly of SiO2@QD supra-NP immunoconjugates was developed. This method is entirely driven by affinity interactions. Functionalization of the SiO2@QDs with imidazole-modified dextran enhanced their colloidal stability, reduced non-specific binding, and was a handle for the assembly of TACs suitable for selective immunolabeling and imaging of breast cancer cells. As the self-assembly method allows for simple preparation of multiple colors of SiO2@QDs, multiplexed immunolabeling and imaging applications are anticipated in the future. The SiO2@QDs also displayed exceptional photostability, which is ideal for cellular imaging, and the brightness of the materials makes them viable for bioanalytical 137  applications where maximizing sensitivity is vital—for example, point-of-care diagnostic applications with smartphone detection. The SiO2@QD supra-NPs are thus promising materials for applications in bioanalysis and imaging.  6.4 Experimental section  6.4.1 Preparation of SiO2@(QD-Dex) conjugates 6.4.1.1 Preparation of API-modified dextran (API-Dex) Scheme 1 summarizes the preparation of API-modified dextran. The protocol is for 10 kDa dextran but is readily adapted to 6 kDa dextran.  Dextran (1, 0.50 g, ~10 kDa MW, ~50 µmol of polymer chains or ~3.1 mmol anhydroglucose) was weighed into a 40 mL glass vial and dissolved in 20 mL deionized water. An aliquot of 0.1 M NaIO4 (3.1 mL, 0.31 mmol, 0.1 equiv.) was added and the sample was covered in foil and mixed at 4 C for 8–12 h (overnight). The samples were then dialyzed against 1.0 L of deionized water using 3.5 kDa-MWCO dialysis tubing for 2  ~24 h periods with one water change in between. The purified samples were freeze-dried to yield a fluffy white solid of oxidized dextran (3, Ox-Dex).   Scheme 1 Synthetic scheme for API-modified dextran. The R group may be another secondary amine linkage to API, an unreacted aldehyde (or the corresponding hydrate), or have cyclized with the secondary amine of the shown API modification. 138  Ox-Dex (3, 0.10 g, 0.617 mmol anhydroglucose, maximum 0.123 mmol aldehyde if 100% efficiency in the preceding step) was added to a 5 mL glass vial and dissolved in 2 mL of deionized water. Neat 1-(3-aminopropyl)imidazole (API, 2; 19.7 µL, 0.165 mmol, 1.3 equiv.) was added and the solution was mixed at room temperature for 2 h. Next, an aliquot of NaCNBH3 (aq) (0.25 mL of 51 mg/mL in deionized water) was added to the reaction, which was then left to mix overnight at room temperature. The final reaction mixture was pipetted into 95% ethanol (20 mL) to precipitate the dextran, which was then pelleted via centrifugation at 13 000 RCF for 5 min. The supernatant was removed, the pellet re-dissolved in 2 mL of deionized water, and re-precipitated with ethanol followed by centrifugation to collect the pellet of API-modified dextran (4, API-Dex). The pellet was dried under reduced pressure to yield an off-white powder.  6.4.1.2 Preparation of supra-NPs IPS-functionalized silica NPs (200 µL) were mixed with GSH-coated QDs (20 µL, 2.4 M) for  1 h at room temperature. The self-assembled SiO2@QD supra-nanoparticles were isolated from excess QDs by centrifugation (14 000 RCF, 10 min) and redispersed in borate buffer (100 µL). After three cycles of precipitation and redispersion, the SiO2@QD supra-NPs were dispersed in 200 µL borate buffer.  For coating SiO2@QD with dextran (for cellular imaging only), the SiO2@QD (40 µL, ~12 nM) were mixed with API-Dex (40 µL, 50 mg/mL) for 1 h at 60 oC. The resulting SiO2@(QD-Dex) were isolated from excess API-Dex by centrifugation (14 000 RCF, 10 min) and redispersed in borate buffer (40 µL).  6.4.2 Immunolabeling of fixed SK-BR3 cells with SiO2@(QD-Dex) The following sub-sections describe the preparation of immunoconjugates and procedures for immunolabeling, imaging, and image processing.  6.4.2.1 Preparation of TAC Anti-HER2 complexes The details for the preparation of TAC anti-HER2 can be found in section 5.4.2.1.1.  139  6.4.2.2 Immunolabeling fixed SK-BR3 cells Fixed SK-BR3 cells were prepared using the method described earlier in section 4.4.1.1  Briefly, 10 µL of 0.5 × 106 cell/mL fixed SK-BR3 cells were pipetted into a 1.7 mL microcentrifuge tube, followed by 5 µL of TAC-anti-HER2 (0.48 pmol, 97 nM) and 5 µL of DAPI (2.9 µM). The cells were incubated with the reagents for 15 min at room temperature in the dark. Following incubation, 2 µL of either Dex-QD605, SiO2@(QD605-GSH), or SiO2@(QD605-Dex) was spiked into the sample and allowed to incubate at room temperature for another 15 min in the dark. A control sample, which did not contain TAC, was prepared as above but was spiked with 5 µL of 1× PBS buffer instead. The samples were pelleted via centrifugation at 55 RCF for 5 min. The supernatant, containing excess and unbound particles was removed. The cells were then resuspended in 25 µL fresh 1× PBS and imaged on a research-grade microscope or the smartphone-based imaging platform.  For multi-color cellular immunolabeling, the above protocol was used, however, the cells were not DAPI stained.  6.4.3 Imaging SiO2@(QD-Dex) immunolabeled SK-BR3 cells The following sub-sections describe methods for the microscopic, and smartphone-based imaging of immunolabeled cells, and methods for image processing.  6.4.3.1 Microscope imaging For imaging cells with a research-grade microscope, the samples were drop cast on to a microscope slide and a cover slip was applied. Imaging was done with an IX83 inverted epifluorescence microscope (Olympus, Richmond Hill, ON, Canada) equipped with an X-Cite 120XL metal-halide light (Excelitas Technologies, Mississauga, ON, Canada), a 405/20 (center line/bandwidth in nm) for excitation, a 590 nm cut-off dichroic mirror, and 600 nm long-pass for emission, and an Orca-Flash 4.0 V2 sCMOS camera (C11440; Hamamatsu Photonics, Hamamatsu, SZK, Japan). ImageJ software was used for processing images.  140  Single-particle imaging was done on an IX83 microscope equipped with an iChrome MLE laser engine (Toptica Photonics AG, Munich, Germany) and Andor iXon ultra 888 EMCCD camera (Andor Oxford instruments, Belfast, UK).   6.4.3.2 Smartphone-based imaging For imaging cells on a smartphone-based imaging platform (see Figure 7.2),250 a labeled suspension of cells (10 µL, ~0.2 × 106 cell/mL, in 1× PBS) was pipetted into a Countess cell imaging chamber slide (Invitrogen, Carlsbad, USA). This slide was then inserted into the smartphone-based imaging platform for imaging.  6.4.3.3 Image processing To calculate contrast ratios, the mean PL intensity of cells incubated with QDs or SiO2@QD with TAC and without TAC was measured using the analyze particle function in ImageJ. The background intensity was subtracted and the average of the mean per-cell PL intensities was then calculated. The contrast ratio was calculated by dividing the average PL intensity with TAC by the average PL intensity of cells without TAC.   Signal-to-background ratios were calculated from the average of the mean per-cell PL intensity using the analyze particle function in ImageJ, subtracted by the background intensity, and then divided by the background intensity.  141   Supraparticle assemblies of magnetic nanoparticles and quantum dots for selective cell isolation and counting on a smartphone-based imaging platform   This chapter is an adaptation Tran, M.V., Susumu, K., Medintz, I.L., Algar, W.R., Supraparticle Assemblies of Magnetic Nanoparticles and Quantum Dots for Selective Cell Isolation and Counting on a Smartphone-Based Imaging Platform, Anal. Chem. 2019, 91, 11963-11971, with permission from the American Chemical Society (Copyright 2019). Refer to the Preface for full details of author contributions. I had the primary role in obtaining all of the data presented in this chapter.   7.1 Introduction  There is a great interest in the detection and enumeration of specific cell types for diagnostic and therapeutic purposes. For example, in the food industry, the detection of microorganisms helps prevent the spread of foodborne illnesses.251 There are also many examples of medical applications for the enumeration of specific cell types: CD4+ cell counts are routinely used to diagnose AIDS and monitor the health of patients infected with HIV;252,253 hematopoietic progenitor cell counts aid in feasibility assessments for autologous transplantation following ablative chemotherapy and radiation therapy;179–181 and (circulating) tumor cell counts correlate with the status of a cancer in terms of aggression, metastasis, or recurrence, and can help guide treatment.254 Once samples are collected, the target cell type is frequently isolated or distinguished from non-target cells through immunolabeling of diagnostic cell-surface antigens. For example, the expression level of HER2 is useful for characterizing breast cancer aggression and the feasibility for therapy with monoclonal antibodies such as Herceptin (Trastuzumab).255 These and many other examples highlight the importance of developing technologies for the enumeration of specific cell types.256,257   Currently, the gold standard technology for enumeration of specific cell types is flow cytometry, which generally relies on fluorescent immunolabeling of cells.16 Flow cytometers are effective at phenotyping (and sometimes sorting) heterogeneous populations of cells, and can provide 142  information about cell size and shape, but their large size, high cost, sophistication, and need for specialized training make them unsuitable for point-of-need applications. Hemocytometers and Coulter counters are other standard cell counting methods.258,259 Although much simpler and far less expensive than flow cytometry, these methods lack the biochemical selectivity and information that comes from fluorescent immunolabeling. Simple, portable, and low-cost technologies that support fluorescent immunolabeling are needed to maximize the prospective benefits of cell counting assays for pathogen detection, health care, and other bioanalyses. To this end, smartphones are an emerging platform for a wide variety of bioanalyses.111,260,261 Their portability, cost, ubiquity, connectivity, and processing capabilities are amenable to point-of-need and low-resource settings.111 In the case of optical bioanalyses, the smartphone camera has supported colorimetric,262–264 fluorescent,124,182,265–267 phosphorescent,268 and holographic269,270 detection, and the range of analytes has included 2,4-dichlorophenoxy acetic acid (a herbicide),262 human cytomegalovirus,182 human alpha-thrombin,124 and human chorionic gonadotropin,268 among many others.  Here, we present a prototype smartphone-based imaging platform (SIP) paired with magnetic-luminescent supraparticle assemblies to enable simple, rapid, and selective cell isolation, fluorescent labeling, and counting. The materials and assay format are outlined in Figure 7.1. The assemblies (denoted MNP@QD) comprise dextran-coated magnetic iron oxide nanoparticles (MNPs) surrounded by a dense corona of semiconductor QDs and are further stabilized with an outer layer of dextran. These assemblies are largely self-assembled through affinity interactions: the dextran on the MNPs is modified with imidazole groups (API) to bind QDs through coordinate interactions, and immunoconjugates of the MNP@QDs are prepared via TACs. One end of the TAC is an anti-dextran antibody that binds to the MNP@QD and the other end of the TAC binds to the target cell type.12 For proof-of-concept, we demonstrate an assay for counting SK-BR3 breast cancer cells on the basis of their expression of HER2, including selectivity against the HER2-negative MDA-MB-231 breast cancer cell line. MNP@QD and TAC are spiked into a sample cell suspension, and, after a brief incubation period, a magnet is applied to pellet the  SK-BR3 cells via bound MNP@QD-TAC conjugates. The pellet is then washed, resuspended, and transferred into a chamber slide for imaging on the SIP. The cells brightly glow the color of the 143  QD PL and cell counts are obtained from the smartphone image. The SIP and MNP@QD materials are both important advances toward point-of-need technologies for cell-based assays.    Figure 7.1 (A) Schematic of an MNP@QD supraparticle assembly: QDs are bound to an imidazole-modified dextran coating on the MNPs and overcoated with additional imidazole-modified dextran. (B) Schematic of TAC-mediated binding of an MNP@QD to HER2 antigen on the surface of an SK-BR3 cell and isolation by magnetic pull-down. (C) Zoomed view of the TAC-mediated binding. (D) Diagram illustrating the steps in the cell counting assay: TAC and MNP@QD are added to a sample cell suspension and the target cells are pelleted magnetically, washed, resuspended, and transferred to a chamber slide for enumeration on the SIP. The assay is demonstrated with a mixture of HER2-positive and HER2-negative breast cancer cells.   7.2 Results and discussion  7.2.1 Smartphone-based imaging platform  Figure 7.2A shows a rendition of the prototype SIP, which was manufactured by 3-D printing. The top of the SIP was a stage with a recess that aligned the smartphone camera with additional 144  optics situated in the lid. This top-stage manually translated along the vertical axis via two spring-loaded thumbscrews as a simple focusing mechanism. The optics, illustrated in Figure 7.2B, included a 500 nm long-pass emission filter and two half-inch lenses (ƒ = 25 mm bi-convex; ƒ = 19 mm achromatic doublet). The optics provided an imaging field-of-view of 10.5  7.9 mm2, corresponding to an image pixel size of ~2.6 µm. Cells were counted in a smaller area of the image, 2 × 3 mm2, where excitation (vide infra) was most uniform and imaging aberrations were at a minimum. The emission filter sat in a holder that slid in and out (without moving the smartphone) for quick changes of filters to match the emitter of interest.     Figure 7.2 (A) Rendering of the smartphone-based fluorescent cell imaging platform (SIP). (B) Optical design for imaging (L1, ƒ = 25 mm plano-convex; L2, ƒ = 19 mm achromatic doublet lens; emission filter). The laser is 20 mW at a wavelength of ~405 nm. The laser beam-shaping optics are omitted for clarity but can be found in Figure 7.3. (C) Photo of the SIP with the smartphone mounted. The thumb-screw stage-adjusters are used for manual focusing. The laser-adjustment dial modifies the angle and lateral position of the laser. The smartphone powers the laser diode, the output intensity of which can be adjusted with the rheostat. (D) Photo of the SIP with top stage removed for a top-down view, with the sample slide illuminated by the laser.    145  The excitation source in the SIP was a 20-mW violet (405 nm) laser diode, which is both a common and ideal excitation wavelength for QDs.28,271 The laser output was shaped through a series of lenses (Figure 7.3) to illuminate a narrow section of the sample from below. A dial translated the laser in the lateral direction and adjusted the angle of the beam. The laser was powered through the microUSB port of the smartphone and its intensity was manually adjusted via a rheostat (see Figure B-1). The sample (as a cell suspension) was loaded into commercially available cell-counting slides that had two plastic chambers, each 17  6 mm2 in area and 0.1 mm high, holding a sample volume of ~10 µL. The volume of the section of the image where cells were counted was ~0.6 µL. Each slide was useful for two cell-counting measurements, which were validated on the SIP by imaging fixed SK-BR3 breast cancer cells stained with 4′,6-diamidino-2-phenylindole (DAPI). This nuclear stain was chosen for validation because it provided high signal-to-noise ratios (10 ± 6, written as average ± one standard deviation, SD) and was adequately excited by the SIP laser diode. Dilutions of fixed cells with concentrations between 5.5 × 102 and 5.5 × 106 cells/mL were counted on the SIP and, in parallel, on a commercial instrument (Countess II Cell Counter). A near 1:1 correlation between the two measured concentrations was observed (see Figure 7.4, slope = 1.01, correlation coefficient R2 = 0.996).    Figure 7.3. Laser diode optics. The dispersed light from the laser diode is focused through an aspheric lens (ƒ = 3.3 mm) and a plano-convex lens (ƒ = 15 mm), then diverged by a plano-concave lens (ƒ = –15 mm).  146    Figure 7.4 (A) Comparison of fixed and DAPI-stained SK-BR3 cell-counting on the SIP and a commercial Countess cell counter. The Countess II cell counter enumerates the cells by imaging the samples in brightfield mode in solution and applying their cell-counting image analysis algorithm. (B) SIP images of DAPI-stained SK-BR3 cells at various concentrations.   147  7.2.2 Composite MNP@QD nanoparticles  7.2.2.1 Necessity of API-modification of MNPs To enable concurrent magnetic isolation and fluorescent labeling of cells, a composite nanoparticle (MNP@QDs) was developed. The dextran coating of commercially available MNPs were chemically modified with API for self-assembly with ligand-stabilized QDs. The QDs were generally coated with either glutathione (GSH) or histidine (His) ligands, although it was possible to assemble QDs coated with other ligands as well (e.g. zwitterionic compact ligand CL4,272 data not shown). The API-modification was necessary for efficiently assembling aqueous QDs to the MNPs. Figure 7.5 shows PL emission images of magnetically-isolated dextran-coated MNPs (i.e. as-received from supplier, Dex-MNP) and API-modified MNPs after mixing for 60 min with GSH-QD605. Integrating the total PL intensity from the nanoparticles in the image field of view (using a spectrometer coupled to the trinocular head of the microscope) indicated that the API modification resulted in ~180-fold brighter nanoparticles. This result was interpreted as ~180-fold more QDs per MNP with the API modification.   Figure 7.5 (A) Comparison of GSH-QD605 binding to API-MNP and MNP. PL emission microscopy of API-MNP and MNP after mixing with GSH-QDs. A pixel intensity bar is shown for reference. Scale bar = 200 µm. The out-of-scale white coloring is for illustrative purposes; the PL signal did not saturate the detector. Quantitative analysis was not affected by this scaling. (B) The measured PL intensity for each sample was measured on a spectrometer that was coupled to the trinocular head of the microscope. The MNP concentrations, as measured by NTA, were kept the same. 148  7.2.2.2 Estimating the number of QDs per MNP Although there is batch-to-batch variation, the number of QDs assembled per API-modified MNP (API-MNP) was estimated to be 1000 ± 400 for QD635 and 9200 ± 3600 for QD575 (Figure 7.6). The number of QDs per MNP was estimated by UV-visible extinction measurements of the composite NPs, bare MNPs, and QDs. A double-beam instrument was necessary to obtain adequate signal-to-noise ratios. The QD contribution to the spectra of the composite NPs was deconvoluted from the MNP scattering contribution, and the concentration of QDs was determined via the Beer-Lambert law and the absorbance at the first-exciton peak. The path length was 1 cm and the extinction coefficients were calculated from previously reported values.273 Figure 7.6 shows the data for MNP@QD575 and MNP@QD635. The range of QDs per MNP@QD assembly was estimated from NTA measurements of the PL intensity distribution (vide infra for Figure 7.13) for QD635 (1 000 ± 400 QDs per MNP) and QD575 (9 200 ± 3 700 QDs per MNP). Note that a pull-down assay, which looked at the absorbance of excess QDs in the supernatant after MNP@QD assembly, suggested values of 12 000 ± 200 QD635 per MNP and 30 000 ± 2 700 QD575 per MNP (data not shown), which we deemed unrealistically high. In contrast, the values derived from Figure 7.6 are consistent with full coverage of the MNP surface area. Assuming the MNPs are spheres with a diameter equal to their hydrodynamic diameter from NTA measurements, the maximum loading of QDs is ~7 100 for QD575 and ~2 500 for QD635. However, the real MNPs are not ideal spheres and QDs may embed within the dextran layer (not just on its surface), increasing the overall loading capacity.   149    Figure 7.6. UV-visible extinction spectra, deconvoluted contributions, and fits for estimating the number of QDs assembled per MNP for (A) QD635 and (B) QD575.    7.2.2.3 Importance of overcoating with API-Dex In a final step, the MNP@QD assemblies were overcoated with API-Dex. MNP@QDs that were not overcoated with API-Dex or mixed with unmodified dextran quickly flocculated out of solution. In contrast, MNP@QD mixed with API-Dex remained as an optically clear solution, indicative of colloidal stability from the API-mediated functionalization (see Figure 7.7). The API-Dex was also useful for conjugation of the cell-targeting antibody (vide infra).   150    Figure 7.7. Colloidal stability of MNP@QD overcoated with API-modified dextran (API-Dex). (A) Pellet of MNP@QDs shortly after initial preparation. (B) Flocculation of MNP@QDs when incubated with unmodified dextran. (C) Colloidally stable MNP@QD overcoated with API-Dex after the same incubation period. (D) As-prepared MNP@QD after pelleting the particles, followed by removing the supernatant, and then washing the pellet with fresh carbonate buffer. The particles were resuspended in fresh buffer and then allowed to incubate on the benchtop for 5 min. MNP@QD incubated with either (E) unmodified dextran or (F) API-Dex. All photographs were taken under ambient lighting conditions. The arrows indicate sample pelleting or flocculation.   7.2.2.4 Additional characterization of MNP@QDs Figure 7.8 shows representative TEM images of API-MNP and MNP@QD. The API-MNP are non-spherical and show high-contrast iron oxide at their core. In contrast, the MNP@QD assemblies show a high density of smaller spherical dark spots across their diameter, indicative of a dense assembly of QDs.  151   Figure 7.8 (A) TEM images of API-MNP and (B) MNP@QD605 (9.8 ± 1.3 nm diameter for QD605). Insets show zoomed views. Inset scale bars are 50 nm. Full image scale bars are 200 nm.    7.2.2.4.1 SEM-EDX Analysis of the API-MNP and MNP@QD by SEM-EDX (Figure 7.9) shows the expected signals for the API-MNP (Fe) and QDs (Cd, Zn, S). Additional data is also provided within Figure 7.9. Briefly, the EDX analysis of the MNP indicated the presence of Fe and O, consistent with the iron oxide composition (Spot 1). Analysis of the MNP@QD indicated the presence of Fe, O, Zn, S, and Cd, consistent with a composite of iron oxide MNP and QDs (Spots 4–7). (Se signals were too weak to observe.) Regions of the images without nanoparticle material (Spots 2 and 3 for API-MNP; Spots 1, 2, 3, and 8 for MNP@QD) did not show these elements.    152    Figure 7.9 (Left) EDX analysis of API-MNP, MNP@QD, and sample substrate (control). Inset graphs represent the 2ꟷ4 keV range. Red-dotted lines align with S and Cd peaks. (Middle, Right) SEM-EDX point analysis of API-MNP and MNP@QD. Spots represent different regions of the SEM images that were analyzed by EDX. In the MNP@QD EDX figure, Spot 4 and 5 were chosen to represent the MNP@QD, and Spots 6 and 7 provided analogous spectra (not shown). Inset graphs represent the 3–4 keV range. Likewise, Spot 1 was chosen to represent the background and substrate, and Spots 2, 3, and 8 provided analogous spectra (not shown).   7.2.2.4.2 X-ray photoelectron spectroscopy X-ray photoelectron spectroscopy (XPS) was used to confirm the EDX results and was done on samples of unmodified MNP, API-MNP, and MNP@QD. As seen in (Figure 7.10), all samples showed characteristic peaks (2p, 3s, 3p) for Fe. Only the MNP@QD sample showed peaks characteristic of the QDs: Cd (3p, 3d, 4p), S (2s, 2p), and Zn (2p). Again, Se signals were too weak to observe.  153    Figure 7.10. XPS elemental analysis of MNP, API-MNP, and MNP@QD (no API-Dex overcoating). Peaks are labelled according to the element. Peaks corresponding to QDs are labelled in red (Zn, Cd, S). Peaks corresponding to MNPs are labelled in black (Fe, O, C). Peaks corresponding to the glass background (Si) and potassium or sodium salts are labelled in grey.   7.2.2.4.3 Infrared spectroscopy The MNP@QD (where the QDs were GSH-QD605) were also characterized by attenuated total reflectance infrared absorption spectroscopy (ATR-IR) and compared to GSH-QDs and API-MNP, as shown in Figure 7.11. The results suggested that QDs retained many of their original ligands upon assembly with the MNPs. The MNP@QD sample showed peaks characteristic of both the GSH-QDs and the API-MNP. These peaks included resonances at 1560 cm–1 and 1345 cm–1 (GSH-QDs and MNP@QD), a strong and broad peak at 550 cmꟷ1 (API-MNP and MNP@QD; indicative of Fe–O vibration modes in the iron oxide), and a strong and sharp resonance at 1010 cm–1 (API-MNP and MNP@QD; a C–O stretching vibration of dextran). As 154  expected, all samples had resonances at ~3300 cm–1 and ~2900 cm–1, corresponding to O–H and C–H stretching vibrations. A water bending peak at 1650 cmꟷ1 was prevalent for the API-MNPs but less so for the MNP@QDs and GSH-QDs. Overall, the data confirmed that the QDs in the MNP@QD assemblies retained many of their GSH ligands upon assembly with the API-modified MNPs.     Figure 7.11. Stacked ATR-IR spectra for API-MNP, MNP@QD, and GSH-QD. Fe-O, C-O, C-H, and O-H bond stretching vibrations are labelled where applicable.    7.2.2.4.4 Nanoparticle tracking analysis (NTA) Nanoparticle tracking analysis (NTA) was used for size analysis of the API-MNP and MNP@QD, using both scattering and PL measurement modes, although only the MNP@QDs were trackable in PL mode. Figure 7.12 shows that the average size of the MNP@QD575 was 250 ± 81 nm 155  (± standard deviation) by scattering and 258 ± 87 nm by PL. These values were similar to the size of the MNP by scattering (252 ± 117 nm), although the size distribution of the API-MNP was broader. The mode size increased from 229 ± 2 nm for the API-MNPs to 238 ± 3 nm (± standard error) for the MNP@QD, consistent with assembly of the QDs. Additional NTA characterization for MNP@QD575, MNP@QD605, and MNP@QD635 are shown in Figure 7.13.    Figure 7.12 (Left) API-MNP and MNP@QD575 size distributions (252 ± 117 nm, 250 ± 81 nm) determined by NTA. The API-MNP had no measurable PL emission. (Middle, right) NTA sizing data for MNP@QD605 and MNP@QD635.    156    Figure 7.13. Plots of intensity versus size (top row; scatter plots) and concentration versus intensity (bottom row; histograms) for MNP@QD575, MNP@QD605, and MNP@QD635.   7.2.2.4.5 MNP@QD and QD optical characterization The UV-visible extinction spectrum of the MNP@QD575 (Figure 7.14A) largely resembled the scattering spectrum of the MNPs; however, the spectrum for the MNP@QD575 had a small feature at ~550 nm, which corresponded to the first exciton peak of the QD575. The PL emission spectrum of the MNP@QD575 was centered at 575 nm with a FWHM of 26 nm. Other than emission spectra shifts < 5 nm, the emission properties of the MNP@QD matched those of the QDs alone for all colors of QD tested. The PL excitation spectra of the MNP@QD were also consistent with the QDs alone (Figure 7.14B-C). The MNPs had no detectable PL emission.  157    Figure 7.14 (A) UV-visible extinction (Ext) and PL emission (Em) spectra for API-MNP and MNP@QD575. The arrow highlights the first exciton peak of the QD575. (B) PL excitation spectra for various colors of MNP@QD. (C) PL excitation spectra for the corresponding His-QD.The QD colors measured were  = 575 nm, 605 nm, and 635 nm.   7.2.2.4.6 Fluorescence microscopy imaging of MNP@QDs Individual MNP@QD assemblies were detectable by fluorescence microscopy, as seen in Figure 7.15, where the single assemblies were distinguished from dimers, trimers, and larger groupings by their measured intensities, which scaled by approximate integer values.     Figure 7.15 Fluorescence microscopy image of MNP@QD575 (left; scale bar = 5 µm). A three-dimensional pixel-intensity plot for the boxed area of the image is provided (right). Single MNP@QD assemblies are indicated by the white arrows.    158  7.2.2.4.7 Smartphone imaging of magnetic and PL MNP@QD properties The combined magnetic and PL characteristics of the MNP@QD are shown in Figure 7.16, where a colloidal solution of MNP@QD575 was quickly pelleted via a permanent magnet and long-wave UV-illumination induced bright yellow PL from the pellet, which was captured via smartphone camera.   Figure 7.16 Magnetic pelleting of MNP@QDs from a colloidal suspension.   7.2.3 Immunomagnetic cell isolation and imaging  Fixed SK-BR3 breast cancer cells were chosen as a model cell line to demonstrate proof-of-concept for immunomagnetic cell isolation and imaging on the SIP. This cell line overexpresses the HER2 antigen and was therefore targeted with an anti-HER2 antibody. The anti-HER2 was conjugated to the MNP@QD as part of a bifunctional TAC with an anti-dextran antibody (see Figure 7.1C). MNP@QDs were prepared in four different PL emission colors (blue, QD485; yellow, QD575; orange, QD605; red, QD635). Fixed SK-BR3 cells were incubated with TAC conjugates of one of the colors of MNP@QD for 5 min, magnetically pelleted, washed, resuspended in PBS buffer, and an aliquot (10 µL) pipetted into a chamber slide. Figure 7.17 shows representative images of the isolated cells acquired with the SIP, and the insets show examples of high-magnification images of the isolated cells acquired from a research-grade microscope. The signal-to-noise ratios (SNRs) on the SIP for the four colors of MNP@QDs were 7 ± 3 (± SD), 8 ± 4, 14 ± 6, 9 ± 4 for cells labeled with MNP@QD485, MNP@QD575, MNP@QD605, and MNP@QD635, respectively. The corresponding SBRs were 6 ± 3, 8 ± 4, 17 ± 8, and 10 ± 5 respectively.  159    Figure 7.17 Smartphone (main images, scale bar = 200 µm) and microscope images (insets, scale bar = 20 µm) of fixed SK-BR3 cells isolated with MNP@QDs of various colors: QD485, QD575, QD605, and QD635, where the notation QDλ refers to the wavelength of peak PL emission for the QD. The smartphone images were acquired in RGB color format. Microscope images are pseudo-colored from the measured monochrome intensity values.    MNP@QD635 composites were compared to other formats for cell isolation and imaging using a research-grade fluorescence microscope. Figure 7.18 shows the average PL intensities for cells and background regions of images, as well as their variation. Unlabeled cells (i.e. imaged via autofluorescence) had a SNR of 3.5 ± 0.3 (± 1 SD). Cells immunolabeled with individual QD-TAC conjugates had a SNR of 9 ± 1, and cells isolated and labeled with a mixture (not composite) of MNP-TAC conjugates and QD-TAC conjugates had a SNR of 16 ± 7. In contrast, the MNP@QD composites provided a vastly superior SNR of 40 ± 17. The corresponding SBRs were 0.3 ± 0.02, 0.9 ± 0.1, 1.4 ± 0.7, and 4.3 ± 2. All imaging was done with the same microscope settings and under the same conditions.   160   Figure 7.18 (A) Differential interference contrast (DIC) and PL images of labeled SK-BR3 cells acquired with a research-grade microscope for different materials combinations. Controls were unlabeled SK-BR3 cells (the cell is outlined with a dashed circle). Scale bars = 20 µm. All images were acquired under the same microscope and camera settings. PL images are false-colored from the measured monochrome intensity values. (B) Average PL signal from labeled cells and surrounding background for each materials combination. A minimum of 20 cells were analyzed from images acquired at a lower magnification (10×). 161  7.2.4 Live cell counting  The MNP@QD composites and SIP were tested for live cell isolation and counting. The same procedure used for fixed cells was repeated with live SK-BR3 cells. Figure 7.19A-B shows representative SIP images of an increasing number of isolated live SK-BR3 cells and a plot of the number of cells counted versus the number of cells spiked. The data was linear (correlation coefficient of R2 = 0.99) with a slope that corresponded to a live-cell capture efficiency of  72 ± 3%. Isolation and counting of live SK-BR3 cells was also assessed in the presence of a constant number (~6100) of background MDA-MB-231 breast cancer cells. MDA-MB-231 cells do not express HER2 and therefore should not have been immunomagnetically isolated. As shown in Figure 7.19C, the linear trend (R2 = 0.97) and the SK-BR3 capture efficiency (68 ± 6%) were maintained with the high number of background MDA-MB-231 cells. As another test of the specificity of SK-BR3 isolation, counting assays were done with a constant number of spiked  SK-BR3 cells and an increasing background of MDA-MB-231 cells. Figure 7.19D shows that spikes of ~4500 (± 500) SK-BR3 cells were successfully counted with recovery between 70–83% against backgrounds of ~800 and ~8600 MDA-MB-231 cells, consistent with the data in Figure 7.18B-C. In a one-tailed t-test, there was no statistically significant difference in the mean number of cells enumerated between each of the samples (determined by a one-way analysis of variance for a significance level of   0.10).   162     Figure 7.19 (A) SIP images of increasing numbers of SK-BR3 cells isolated with MNP@QDs. A single cell is indicated with the arrow for the 100-cell sample. Scale bar = 200 µm. Each image represents an area of 2 × 3 mm2. (B) Quantification of an increasing number of live SK-BR3 cells (HER2+). (C) Quantification of an increasing number of live SK-BR3 cells in samples spiked with a background of 6140 ± 470 live MDA-MB-231 cells (HER2ꟷ). (D) Quantification of a constant level of live SK-BR3 cells with increasing background of live MDA-MB-231 (HER2ꟷ) cells. All samples were in separation buffer. Additional smartphone images for the assays in panels B-D are shown in Figure B-2, Figure B-3, Figure B-4.   7.2.5 Discussion  In its present and relatively unoptimized format, our assay can quantify cells between a lower limit of ~102–103 cells/mL and an upper limit of ~107 cells/mL. The upper limit of this range is set by the requirement (for the counting algorithm) for dark space between individual cells in SIP PL images. The lower limit is set by the stochastic probability of finding an individual cell within the 0.6 µL volume of the counting field of view (~1700 cells/mL) or within the 10 µL volume of the chamber slide that can be searched for an individual cell (~100 cells/mL). It is also critical to note that these limits are the cell concentrations after magnetic isolation. The cell concentrations in the original sample can be much lower, provided that the magnetic isolation is efficient, and the magnetic isolate is resuspended in a much smaller volume than the original sample. For this proof-of-concept study, the volume ratio between our samples and resuspended magnet isolates was 1:2, 163  suggesting that there is significant room for magnetic pre-concentration with larger sample volumes (e.g. a 2 mL-sample could be magnetically concentrated 100-fold by resuspension in 20 µL). Our assay was also simple, consisting of only four steps (incubation, magnetic pelleting, single wash, and imaging) with a total assay time under 15 min.  The SIP prototype was simple in design, amenable to manufacturing, and potentially adaptable to many models of smartphone (by varying the precise dimensions of the top-stage). It also had a small footprint (20  10 cm2) and was approximately one third of the cost of the smartphone. In principle, the smartphone could be replaced with a simple CMOS imaging sensor, but this approach would lack the processing, memory, connectivity, and other versatility advantages of a smartphone. The chamber slide sample holder was the same dimensions as a standard microscope slide (25  75 mm2), such that many other sample methods and devices (e.g. microfluidic chips) are potentially compatible with the SIP or straightforward modifications thereof. Although our prototype had the laser controller separate from the SIP housing, future designs can easily integrate the circuit, and an app for measurement control and data analysis can be developed. Future designs could also attempt greater simplicity by evaluating the possibility of using blue light filtered from the camera flash for excitation of PL, as we have shown for other applications.41,96 Nevertheless, the violet laser is advantageous in that it offers greater intensity, is more readily focused on a region of interest, and permits use of all three RGB color channels.   For the SIP, a useful point of comparison is the FDA-approved CellSearch system for circulating tumor cell (CTC) isolation and enumeration. Compared to the SIP, the CellSearch Analyzer is 100-times larger by volume and 20-times larger by footprint, costs approximately two orders of magnitude more, and separates magnetic isolation and fluorescent staining into multiple steps. The SIP and MNP@QD have potentially significant advantages over this technology in terms of size, cost, and combined magnetic isolation and fluorescent staining (albeit currently for only a single biomarker), such that there is strong potential for the enumeration of CTCs as a future application. A summary comparison of the SIP-MNP@QD platform versus other research-and-development-stage smartphone-based cell counting platforms can be found in Table B-1 within the  Appendix B  .132,274,275 Overall, the SIP-MNP@QD platform compares favorably, with the shortest analysis time, smallest on-device sample volume, and a competitive detection limit. 164  Although engineering ingenuity has made several forms of imaging and bioanalysis possible with smartphones (see refs.111,260,261 for reviews), smartphone cameras are neither designed nor optimized for PL imaging. This fact motivated the use of an app (Camera FV-5) that enabled control over ISO and shutter speed (i.e. exposure time) settings, and supported imaging in RAW format to avoid challenges from auto-correction of images. More importantly, the non-optimal PL imaging with a smartphone motivated the design of the MNP@QD assemblies. The assembly of hundreds to thousands of QDs per MNP resulted in ultra-bright nanoparticles that compensated for any sensitivity deficiency of the smartphone camera and the simple optics in the SIP. The combined magnetic and bright PL properties within a single vector represented another simplifying aspect of the supraparticle assembly. Separate magnetic and luminescent materials were not required and, even though unbound MNP@QDs were magnetically isolated in parallel with cell-bound MNP@QDs, isolated cells were visible above background PL with high SNR. This result is attributed to the cell concentrating many MNP@QD assemblies, and thus a very large number of QDs, into a relatively small number of image pixels.  Another advantage of our MNP@QD materials is that they were largely self-assembled. The stock materials are ligand-coated QDs, API-modified MNPs, API-modified dextran, and TAC reagents. Once these materials are on hand, immunoconjugates of MNP@QD can be prepared without covalent chemistry, with assembly driven instead by affinity interactions. The exclusion of covalent chemistry necessitates fewer steps in the process, simpler purification, and makes it more reproducible and amenable to scale-up. It also potentiates on-demand preparation of MNP@QD immunoconjugates using whatever combination of cell-targeting antibody and QD color is desired. Use of the TAC is also advantageous because it puts MNP@QD-conjugated antibodies in a productive orientation for binding to their antigen targets, which is a frequent and significant challenge with covalent conjugation of antibodies to nanoparticles.63   In addition to their self-assembled preparation, our MNP@QD materials are advantageous in their size and in the intrinsic optical properties of the QDs. Compared to larger magnetic-fluorescent particles (e.g. refs.276–278), smaller particles are less prone to aggregating and clumping isolated cells, which would hinder counting, and also block less of the cell surface, which is a benefit to applications in which multiple antigens are to be measured. Compared to magnetic-fluorescent 165  materials based on organic dyes (e.g. refs.279,280), the QDs are expected to be brighter, more resistant to photobleaching, and better suited to multicolor analyses. The cadmium-based QDs used in this study may also be substituted by any other ZnS-shelled heavy metal-free QD material, such as InP/ZnS,281,282 which is of less concern for disposal in practical application. The brightness of the MNP@QD materials also enabled, to our knowledge, this first example of a magneto-immunofluorescent cell counting assay on a smartphone. Previous reports of smartphone-based cell counting have used magnetophoresis with non-specific fluorescent staining,275 immunofluorescent labeling with surface-immobilized capture antibodies rather than magnetic isolation,283 or non-specific staining without magnetism or immunocapture.284 Although each approach has its advantages, our all-in-one magnetic isolation and immunofluorescent staining requires only one antibody against the target cell and does not require a fluidic system.   Moving forward, the MNP@QD materials and SIP hold promise for a simple and effective method for multiplexed isolation, counting, and immunoprofiling of cells. For example, the preparation of MNP@QD with different colors of QDs (e.g. blue, green, yellow, orange, or red emitters) paired with different targeting antibodies is potentially useful for encoding the phenotypes of isolated cells based on their expression of cancer-relevant antigens such as Mucin 1 (MUC1), epithelial cell adhesion molecule (EpCAM), epidermal growth factor receptor (EGFR), estrogen receptor (ER), and progesterone receptor (PR). The same concept can also be applied beyond cancer; for example, assays of hematopoietic progenitor cells or immune cells (e.g. HIV/AIDS), multiplexed screening assays for pathogenic microorganisms, and potentially magnetic pull-down immunoassays for cell-free biomarkers (e.g. proteins, genes, small molecules in bulk solution). The spectrally narrow PL emission of QDs and the efficient parallel excitation of multiple colors of QDs at a single wavelength make QDs particularly well-suited to these types of applications.   7.3 Conclusion   We have presented a design for a smartphone-based imaging platform that, in tandem with MNP@QD supraparticle assemblies, enables selective cell isolation and quantification on a smartphone. The MNP@QD are multifunctional from the standpoint of the magnetism of the MNPs and the PL properties of the QDs. The preparation of these supraparticle assemblies and 166  their immunoconjugates via self-assembly is highly advantageous, and their ultra-bright PL enables imaging of single cells with high SNR. Proof-of-concept was demonstrated through selective isolation and counting of HER2-positive SK-BR3 breast cancer cells against a background of HER2-negative MBA-MD-231 breast cancer cells. We envision future use of the SIP and MNP@QD materials for applications such as pathogen detection, enumeration of cells for medical diagnostics (e.g. immune cells, CTCs), and immunofluorescent profiling of cells, particularly in the context of point-of-need analyses.  7.4 Experimental section  7.4.1 Materials  Dextran RapidSpheres (dextran-coated magnetic iron oxide nanoparticles; MNPs) and the Do-It-Yourself Positive Selection Kit II (for formation of TACs) were from STEMCELL Technologies (Vancouver, BC, Canada). Anti-HER2 antibody (NBP2-32863) was from Novus Biologicals (Centennial, CO). Carbon-coated copper grids (300 mesh) were from Ted Pella (Redding, CA). CdSe/CdS/ZnS QDs (QD575, QD605, QD635) were synthesized using standard methods.285,286 CdZnSe/CdZnS/ZnS QDs (QD485) were synthesized using a published method.287 The notations are of the form QDλ, where λ is the wavelength of peak PL emission. Borate buffer was 50 mM, pH 9.3. Separation buffer was PBS buffer supplemented with 2.0% v/v fetal bovine serum and 1.0 mM EDTA. Dextran (Leuconostoc mesenteroides, 9000–11000 Da MW or Leuconostoc spp. ~6000 Da MW) (6 kDa dextran was used only for the experiments in Figure 7.17), sodium (meta)periodate, 1-(3-aminopropyl)imidazole (API), sodium cyanoborohydride, epichlorohydrin, sodium borohydride, L-Histidine (His), tetramethyl ammonium hydroxide (TMAH), and reduced L-Glutathione (GSH) were from Sigma-Aldrich (Oakville, ON, Canada). Boric acid and ethylenediaminetetraacetic acid (EDTA) were from Fisher Scientific (Toronto, ON, Canada). Sodium tetraborate decahydrate, potassium carbonate, and sodium bicarbonate were from Amresco (Solon, OH). Dialysis membrane (3.5 kDa MWCO) was from Spectrum Laboratories (Rancho Dominguez, CA). Deionized water was from a Milli-Q Synthesis water purification system (Millipore, Burlington, MA). Borate buffer was 50 mM at pH 9.3.  167  7.4.2 Smartphone-based imaging platform  The smartphone imaging platform was designed using AutoCAD 2017 AutoDesk Student 3-D drafting software (AutoDesk, San Rafael, CA), and the components were 3-D printed on a 5th generation MakerBot Replicator 3-D Printer (MakerBot, Brooklyn, NY) using black PLA filament (MakerBot). A Galaxy S7 smartphone (Samsung, Suwon, South Korea) and the Camera FV-5 Pro app (Version 3.31.4; FGAE Studios, Stuttgart, Germany) were used for imaging. A D405-20 laser diode (405nm, 20mW, 5V 75mA Radial, Can, 3 Lead, 5.6mm, TO-18) from US-Lasers Inc. (Baldwin Park, CA) was used as the excitation source. Cell counting chamber slides (Countess, C10283) were from Invitrogen (Carlsbad, CA). Optical components were from Thorlabs Inc. (Newton, NJ). For imaging, the smartphone ISO setting was 100 and the exposure time ranged from 1/3 to 1/10 s.   7.4.3 Preparation of API-modified dextran (API-Dex) The preparation of API-Dex was described earlier in section 6.4.1.1.  7.4.4 Preparation of MNP@QD.  The following paragraphs detail the preparation of the MNP@QD assemblies.   7.4.4.1 Epichlorohydrin-crosslinked MNP (X-MNP) Dextran-coated magnetic nanoparticles (MNP, 2.8 nM in water; 800 µL) were pipetted into a 5 mL glass vial. The MNPs were magnetically pelleted and the supernatant was removed by pipette. A 1.28 mL solution of 7.7% (v/v) epichlorohydrin was prepared in 1 M NaOH and the pelleted MNPs were resuspended in the epichlorohydrin solution. The vial was sealed and kept in the dark as it incubated for 24 h on a shaker at room temperature. The resulting X-MNP were pelleted magnetically, and the epichlorohydrin solution was removed via pipette. The X-MNP were resuspended in 1 mL of deionized water, subsequently pelleted, and the supernatant removed. The resuspension and pelleting steps were repeated for a total of five washes to remove excess epichlorohydrin. After the final wash, the X-MNPs were resuspended in 1.28 mL of deionized water and used in the next step.  168  7.4.4.2 Oxidized X-MNP (Ox-MNP) A 125 mM solution of NaIO4 (aq) was prepared by dissolving ~27 mg NaIO4 in 1.0 mL deionized water. An aliquot (10.24 µL) of NaIO4 (aq) was added to 1.28 mL of X-MNP in water (final concentration of ~1.0 mM NaIO4). The sample was mixed via pipette and kept sealed and protected from light while mixing on a shaker at room temperature for 1 h. The Ox-MNP were pelleted and washed with 3 × 1 mL volumes of borate buffer. After the final wash, the Ox-MNP were resuspended in 1.28 mL of borate buffer and were used in the next step.  7.4.4.3 1-(3-aminopropyl)imidazole-modified MNP (API-MNP) Ox-MNP were pelleted magnetically and then resuspended in 1.28 mL of a 100 mM solution of 1-(3-aminopropyl)imidazole (API) in borate buffer. The sample was sealed and kept in the dark as it incubated overnight at 4 ºC. Next, 64 µL of 0.25 M (10 mg/mL) NaBH4 in 0.5 M sodium bicarbonate (pH 10.5) was added to the reaction mixture. The sample was mixed in the dark for 20 min. Following mixing, the API-MNPs were pelleted magnetically and washed with 5 × 1 mL volumes of borate buffer. After the final wash, the pellet was resuspended in 1.28 mL of borate buffer.   7.4.4.4 Ligand exchange of QDs L-Histidine-coated QDs (His-QDs) were prepared by weighing 100 mg of L-histidine (0.645 mmol) into a 1.7-mL microcentrifuge tube. The L-histidine was dissolved in 300 µL tetramethylammonium hydroxide solution (TMAH; 25% w/w in methanol). As-synthesized hydrophobic QDs in organic solvent (20 µL aliquot; concentration on the order of 102 µM) were transferred into a second 1.7-mL microcentrifuge tube, then diluted to 900 µL with CHCl3. The L-histidine solution was then added to the hydrophobic QD solution and the sample was mixed by vortex. The sample was kept in the dark and incubated at room temperature for 1 h. Following incubation, a biphasic extraction was performed by adding 200 µL of borate buffer to the reaction mixture, into which the His-QDs dispersed. The aqueous phase was isolated and excess TMAH was removed by precipitating the QDs with absolute ethanol (~600 µL) and centrifugation at 4800 RCF for 5 min. The supernatant was then removed and the His-QDs were redispersed in 100 µL of borate buffer. After three cycles of precipitation and washing, the His-QDs were dispersed in 200 µL borate buffer and stored at 4 ºC until needed. 169  Ligand exchange with glutathione (GSH; to prepare GSH-QDs) was done analogously to the ligand exchange with L-histidine. Ligand exchange with CL4 was done as described previously.272,287   7.4.4.5 Self-assembly of MNP@QD API-MNPs (450 pM, 20 µL in borate buffer) were pipetted into a 1.7 mL-microcentrifuge tube and QDs (20 µL, 29 µM) were added. (The QDs were His-QDs, except for the experiments in Figures 7.5B, 7.8-7.11 which were GSH-QDs, and Figure 7.17, which were CL4-QD485.) The samples were sonicated in a water bath at 30 ˚ C for 15 min. After sonication, the MNP@QDs were pelleted magnetically over a period of 2–3 min. Excess QDs in the supernatant were then removed via pipette and the pellet was washed with carbonate buffer (100 µL, 0.1 M, pH 9.3). The MNP@QD pellet was then resuspended in 20 µL of carbonate buffer and sonicated briefly to break up any large aggregates.  7.4.4.6 Overcoating MNP@QD with API-modified Dextran  MNP@QDs were overcoated with API-modified dextran (API-Dex; vide supra) to further stabilize the assemblies and to provide a handle for conjugation with TAC. A solution of API-Dex (25 mg/mL) was prepared in 0.1 M carbonate buffer (pH 9.3) and a 100 µL volume was transferred into a 1.7 mL microcentrifuge tube. MNP@QD (450 pM, 20 µL) were then added to this solution. The sample was briefly vortexed and sonicated, then incubated at 60 ˚ C for 15 min. The overcoated MNP@QD were then pelleted magnetically over a period of 2–3 min, and the supernatant was removed via pipette. The pellet was washed with carbonate buffer (100 µL) and then resuspended in carbonate buffer (40 µL).  7.4.5 MNP@QD characterization  Transmission electron microscope (TEM) images were acquired using a Hitachi High-Technologies H7600 (Tokyo, Japan). Samples were imaged at an accelerating voltage of 100 kV and were prepared by drop-casting 2 × 0.5 µL of ~1 pM solution of nanoparticles (diluted in deionized water) on a TEM grid with drying in a dark box overnight. Scanning electron microscope imaging and energy-dispersive X-ray (SEM-EDX) analysis was done on a Helios NanoLab 650 Focused Ion Beam system (FEI/Thermo Fisher Scientific, Hillsboro, OR). The samples were 170  prepared by drop-casting on TEM grids. Imaging was done at 2 kV/50 pA for secondary electron mode, and 2 kV/0.20 nA for backscattered electron mode. An accelerating voltage of either 5 kV or 10 kV was used for EDX analysis.  X-ray photoelectron spectroscopy (XPS) was done on a Leybold MAX200 XPS spectrometer (Cologne, Germany) via a survey scan for binding energies between 0 to 1075 eV. The samples were prepared by drop-casting onto acid-washed circular glass slides and allowed to dry overnight in the dark. Sample volumes and concentrations: MNP (5 µL, 1.79 nM), API-MNP (20 µL, 331 pM), MNP@QD605 (15 µL, 331 pM, no API-Dex overcoating).  Attenuated total reflectance infrared spectroscopy (ATR-IR) was done on a Perkin Elmer Frontier FT-IR Spectrometer (Waltham, USA) with a ZnSe ATR crystal. Transmittance was measured between 525 and 4000 cmꟷ1 with averaging over 8 scans. MNP, API-MNP, and MNP@QD605 (post-purification) samples were prepared by adding absolute ethanol (via micropipette) to 10 µL of nanoparticles until the samples flocculated. The flocculated material was then pelleted via magnet, the supernatant was removed, and the samples were resuspended in 30 µL of dichloromethane. Samples were pipetted directly onto the ATR-IR crystal and dried in air before taking measurements. There were no obvious indications of interference from light scattering. The ATR-IR crystal by itself was used as the blank.   The hydrodynamic size and concentration of the MNP and MNP@QD samples were determined on a NS300 Nanoparticle Tracking Analyzer (Malvern Instruments, Malvern, UK) instrument equipped with a 488 nm peak wavelength laser, operating at a maximum power of 45 mW. Measurements were done in both scattering mode and fluorescence mode (500 nm longpass filter). Samples were prepared by diluting 200–500 fold in deionized water to a total volume of ~1 mL.  7.4.6 TAC anti-HER2 complexes  The details for the preparation of TAC anti-HER2 can be found in section 5.4.2.1.1.  171  7.4.7 Cell-counting assay  For live-cell counting assays, cultured SK-BR3 cells (HER2 positive) were harvested, counted using a commercial instrument (Countess II Cell Counter; Invitrogen), then diluted and spiked into separation buffer at the desired concentration. MDA-MB-231 cells (HER2 negative) were similarly spiked into select samples alongside the SK-BR3 cells. TAC with anti-HER2 and MNP@QD were added in sequence to the cell suspension, which then stood at room temperature for 5 min. A permanent magnet was applied to collect the MNP@QD and any bound cells as a pellet. The supernatant was removed, the pellet washed once with separation buffer, and the pellet resuspended in separation buffer. A 10 µL aliquot was then transferred to a chamber slide for imaging on the SIP. The cell-counting image analysis algorithm is described in detail in  Appendix B.3.   7.4.8 Fluorescence microscopy Some characterization experiments made use of imaging on a research-grade fluorescence microscope. This microscope was an IX83 inverted microscope (Olympus, Richmond Hill, ON, Canada) equipped with an X-Cite 120XL metal-halide light source (Excelitas Technologies, Mississauga, ON, Canada), a white-LED transmitted light source, an Orca-Flash 4.0 V2 sCMOS camera (C11440; Hamamatsu Photonics, Hamamatsu, SZK, Japan), motorized filter wheels (Sutter Instruments, Novato, CA), and MetaMorph/MetaFluor software (Molecular Devices, Sunnyvale, CA). Filters and dichroic mirrors were from Chroma (Bellows Falls, VT).   7.4.9 Cell culture SK-BR3 cells (ATCC HTB-30 Manassas, VA, USA), a human breast cancer cell line, were cultured in a humidified incubator with 95% air/5% CO2 at 37 °C. The culture medium was McCoy’s 5A (GE Healthcare, Chicago, IL) supplemented with 10% v/v fetal bovine serum and 1× antibiotic and antimycotic (ThermoFisher, Waltham, MA). Cells were cultured in T25 flasks and sub-cultured every 5–7 days.   MDA-MB-231 cells (ATCC HTB-26 Manassas, VA, USA), a human epithelial breast cancer cell line, were cultured analogously to the SK-BR3 cells; however, the culture medium was Dulbecco's Modified Eagle Medium (DMEM) (Sigma Aldrich) supplemented with 10% v/v fetal bovine 172  serum, 1× antibiotic and antimycotic (ThermoFisher), 2 mM L-glutamine (Gibco, 25030081), and 0.1 mM MEM non-essential amino acids (Gibco, 11140-050).  7.4.10 PBS buffers PBS buffer was from Gibco Life Technologies. The composition of the primary PBS buffer used was pH 7.2, 1.54 mM KH2PO4, 2.71 mM Na2HPO4, 155 mM NaCl, and is denoted as PBS. The primary buffer used for live-cell studies was the above PBS buffer supplemented with 2% v/v fetal bovine serum (Sigma Aldrich) and 1 mM EDTA and is denoted as separation buffer.   7.4.11 Counting DAPI-stained cells on the SIP Cultured SK-BR3 cells were trypsinized, centrifuged, and then fixed in ice-cold ethanol for 15 min. The cells were pelleted via centrifugation at 85 RCF for 5 min and the supernatant was removed. The cells were resuspended in fresh PBS buffer and rehydrated for 15 min. The cells were centrifuged at 85 RCF for 5 min and the supernatant was discarded. The cell pellet was resuspended in a solution of 4′,6-diamidino-2-phenylindole (DAPI; 2 mL, 2.9 µM) and incubated in the dark for 10 min at room temperature. Following incubation, the cells were pelleted via centrifugation, and the supernatant was removed. The cell pellet was then resuspended in 2 mL of fresh PBS buffer. The average concentration of cells (measured in triplicate) using a Countess II cell counter was 5.5 × 106 cells/mL. Triplicate samples with dilution factors ranging from 1.4-fold to 10 000-fold were prepared in 1.7 mL microcentrifuge tubes by dilution with fresh PBS buffer. Each sample was pipetted into a chamber slide and imaged on the SIP. The acquired images were further processed on ImageJ and the cells were counted. The same samples were also counted on the Countess II cell counter (vide supra) for validation. This procedure was used to generate the data in Figure 7.4.        173  7.4.12 Counting and imaging fixed SK-BR3 cells labeled with MNP@QDs This procedure was used to generate the data in Figure 7.17   A suspension of freshly trypsinized SK-BR3 cells (~106 cells) was pelleted by centrifugation at  55 RCF for 5 min. The supernatant was removed, and the pellet resuspended in 2 mL of PBS buffer. A volume of 2 mL of 4% (w/v) paraformaldehyde in PBS was added and the sample gently mixed via pipette. The sample was incubated at room temperature for 5–10 min before pelleting via centrifugation at 55 RCF for 5 min. The supernatant was discarded, and the pellet resuspended in 4 mL of PBS buffer.   For each color of MNP@QD, 10 µL of paraformaldehyde-fixed SK-BR3 cells (in PBS) was pipetted into a 1.7 mL microcentrifuge tube. The SK-BR3 cell suspension was then spiked with pre-formed TAC with anti-HER2 (1 µL, 81 fmol), followed by MNP@QD (2 µL, 1.3 fmol, ~8×108 assemblies). The sample mixture was mixed briefly via pipette and then incubated on the benchtop for 5 min. Following incubation, the labeled cells were pelleted magnetically, and the supernatant removed by pipette. The cells were resuspended in 20 µL of fresh PBS buffer.   For imaging on the SIP, 10 µL of the labeled cell suspension was pipetted into a chamber slide. For research-grade fluorescence microscope imaging, samples were prepared by pipetting 7.5 µL of a suspension of labeled cells onto a microscope slide. A cover slip was applied, and then the sample was inverted and imaged through the cover slip. The fluorescence filter sets used were as listed in Table 7.1. The emission spectra of the cells labeled with MNP@QD485 and MNP@QD575 were acquired with a diode-array spectrometer (Greenwave 16 VIS-50; StellarNet, Tampa, FL) that was coupled to the trinocular head of the microscope via a fiber-optic cable. ImageJ software was used for processing images.       174  Table 7.1 Fluorescence microscopy optics for SK-BR3 labeling. Label Ex. Filter a Em. Filter a, b Dichroic Mirror c Objective Lens d MNP@QD485 405/20 BP 460/50 BP* T425 100 XO MNP@QD575 405/20 BP 550 LP T510 100 XO MNP@QD605 405/20 BP 600 LP T590 100 XO MNP@QD635 405/20 BP 600 LP T590 60 X Notes: a Center wavelength/bandwidth, BP = bandpass filter. b LP ═ longpass filter. c T = transmission cut-on wavelength. All numbers in units of nanometers. d X = magnification factor, air-immersion; XO = magnification factor, oil-immersion.   7.4.13 Counting live SK-BR3 cells on the SIP The following paragraphs describe the procedures used to generate the data in Figure 7.19. For all counting assays, the mean numbers of cells spiked into samples were estimated from triplicate measurements of 10-µL aliquots of each stock cell suspension using Countess Chamber Slides and a Countess II Cell Counter as per the manufacturer’s protocol.   7.4.13.1 Counting increasing numbers of SK-BR3 cells without MDA-MB-231 cells The following procedure was used to generate the data in Figure 7.19B. Cultured SK-BR3 cells were trypsinized, centrifuged, and resuspended in separation buffer. Three stock suspensions of SK-BR3 cells with concentrations of ~8 700, ~100 200, and ~570 300 cells/mL were prepared in separation buffer. For validation and determination of recovery percentages, the number of cells per sample was measured by pipetting an aliquot (10 µL) into a chamber slide for counting on a Countess II cell counter.   Samples were prepared in triplicate with the following numbers of SK-BR3 cells spiked into 1.7 mL microcentrifuge tubes as 10 µL aliquots: ~90, ~500, ~750, ~1000, ~2850, ~5700, and ~57 000 cells. The 90-cell sample was prepared from the 8700-cell stock suspension. The 500-, 750-, and 1000-cell samples were prepared from the 100 200-cell stock suspension. The 2850-, 5700-, and 57 000-cell samples were prepared from the 570 300-cell stock suspension.   Next, a spike of TAC with anti-HER2 (1 µL, 81 fmol) was added to the samples, followed by a spike of MNP@QD605 (2 µL, 1.3 fmol, ~8 × 108 assemblies). The samples were mixed via pipette 175  and then left on the benchtop at room temperature for 5 min. The labeled cells were pelleted magnetically, and the supernatant was removed. The cell pellet was washed with separation buffer (20 µL), and the cells were resuspended in fresh separation buffer (20 µL). An aliquot of each sample (10 µL) was then pipetted into a chamber slide and imaged on the SIP.   7.4.13.2 Counting increasing numbers of SK-BR3 cells with a constant number of MDA-MB-231 cells The following procedure was used to generate the data in Figure 7.19C. Separately, SK-BR3 and MDA-MB-231 cells were trypsinized, centrifuged, and resuspended in separation buffer. The MDA-MB-231 cell suspension was diluted to a concentration of ~1 230 000 cells/mL. Three stock suspensions of SK-BR3 containing mean concentrations of ~16 700, ~167 000, ~484 000, and ~1 180 000 cells/mL were prepared in separation buffer. For validation and determination of recovery percentages, the number of cells per sample was measured by pipetting an aliquot (10 µL) into a chamber slide for counting on a Countess II cell counter.   Samples were prepared in triplicate with the following numbers of SK-BR3 cells spiked into 1.7 mL microcentrifuge tubes as 10 µL aliquots: ~167, ~1 670, ~4 840, ~5 890, and ~11 800. Next, ~6 140 MDA-MB-231 cells were added into each sample. The 167-cell spiked sample was prepared from the 16 733 cell/mL stock suspension. The 1 673-cell sample was prepared from the 167 333 cell/mL stock suspension. The 4 840-cell sample was prepared from the 484 000 cell/mL stock suspension. The 5 892- and 11 783-cell samples were prepared with the 1 180 000 cell/mL stock suspension.  Next, a spike of TAC with anti-HER2 (1 µL, 81 fmol) was added to the samples, followed by a spike of MNP@QD635 (2 µL, 1.3 fmol, ~8×108 assemblies). The samples were mixed via pipette and then left on the benchtop at room temperature for 5 min. Following incubation, the labeled cells were pelleted magnetically, and the supernatant was removed. The cell pellet was washed with separation buffer (20 µL) and the cells were resuspended in fresh separation buffer (20 µL). An aliquot (10 µL) of each sample was then pipetted into a chamber slide and imaged on the SIP.   176  7.4.13.3 Counting a constant number of SK-BR3 cells with an increasing number of MDA-MB-231 cells The following procedure was used to generate the data in Figure 7.19D. Separately, SK-BR3 and MDA-MB-231 cells were trypsinized, centrifuged, and resuspended in separation buffer.  Three samples were prepared in triplicate by pipetting 5 µL of ~895 000 cell/mL stock of SK-BR3 cells into 1.7 mL Eppendorf tubes. The samples were then spiked with 10 µL of separation buffer, 858 000 cell/mL MDA-MB-231 stock, or 84 000 cell/mL MDA-MB-231 stock.   Next, a spike of TAC with anti-HER2 (1 µL, 81 fmol) was added to the samples, followed by a spike of MNP@QD575 (2 µL, 1.3 fmol, 8×108 assemblies). The samples were mixed via pipette and then incubated on the benchtop at room temperature for 5 min. Following incubation, the labeled cells were pelleted magnetically and the supernatant containing unlabeled cells (MDA-MB-231), was removed. The cell pellet was then washed with 20 µL separation buffer and the cells were resuspended in 20 µL of fresh separation buffer. An aliquot (10 µL) of each sample was then pipetted into a chamber slide and imaged on the SIP.    177   Affinity immobilization of semiconducting quantum dots and metal nanoparticles on cellulose paper substrates  This chapter is an adaptation Kim, H.; Tran, M.V., Petryayeva, E.; Solodova, O.; Susumu, K., Medintz, I.L., Algar, W.R., Affinity Immobilization of Semiconductor Quantum Dots and Metal Nanoparticles on Cellulose Paper Substrates, submitted to ACS Appl. Mater. Inter. Refer to the Preface for full details of author contributions. Unless indicated in figure captions, I had a sole or significant role in obtaining the data presented in this chapter.   8.1 Introduction  Nanoparticles (NP) constitute a diverse array of inorganic and organic materials which have at least one dimension between 1–100 nm. The emergent properties that are specific to their nanometer size regime are of interest and are typically not observed in their bulk or molecular counterparts. Two examples of commonly utilized nanoparticles include semiconductor QDs 36,288 and gold NPs (Au NPs).79,289 QDs are quantum-confined materials that have a peak PL wavelength that scales with the size and composition of the QD. The QD PL excitation spectra is broad, and the emission spectra is narrow, which are properties that allows for simultaneous excitation of multiple colors of QDs. Furthermore, QDs are more resistant to photobleaching in comparison to other organic fluorescent dye materials. As such, QDs have been widely utilized in biological applications such as imaging, assays, and sensors,38,39 in addition to optoelectronics, photovoltaics, and LEDs.290–292 Au NPs exhibit localized surface plasmon resonances (LSPRs) that are size- and shape-dependent, large surface area-to-volume ratios, and strong light scattering properties.293–295 Au NPs are versatile materials that have been used in a variety of applications, including catalysis296 and biological imaging and sensing82. Examples of the latter include surface enhanced Raman spectroscopy (SERS), photothermal therapy, dark-field imaging, colorimetric assays, and fluorescence quenching assays.293–295 In each of these circumstances the combination of QDs or Au NPs with other functional materials can enable or provide benefits for many applications.   178  Paper has developed as a material that is valuable and versatile for chemistry. Paper consists of a three-dimensional network of cellulose fibers that are typically processed into sheets. The properties of paper that make it attractive from a chemistry standpoint include low cost, biodegradability, high surface area-to-volume ratio, and robustness, and malleability. Furthermore, paper exhibits capillary forces when in contact with fluid, and there are many available methods for chemically functionalizing the cellulose fibers.297–299 The intrinsic capillary action exhibited by paper have been utilized in many microfluidic devices as the driving force for devices suited for bioanalytical assays, point-of-care diagnostics, and for use in low-resource settings.300–302 Many researchers have also utilized paper as a platform in electronic devices,303 and as a scaffold for other functional materials, such as nanoparticles.304 Platforms that have integrated paper substrates and Au NPs or QDs have been developed for numerous applications including a paper- and QD-based protease sensor with smartphone imaging;124,305 glucose assays;306 and DNA hybridization assays based on FRET.307 Paper devices that have integrated Au NPs have been used in SERS studies;308–312 electrodes,313 and as supported catalysts.314  In this chapter, we demonstrate three chemical modifications of cellulose paper substrates for the affinity immobilization of QDs and both Au and platinum (Pt) NPs. The synthetic design was inspired by methods that are traditionally employed for rendering nanoparticles hydrophilic with ligand- and polymer-based coatings. The paper substrates were chemically functionalized with imidazole or thiol moieties that form dative coordinate bonds with the metal surfaces of the nanoparticles. The molecular changes in the modified paper substrates were characterized by  x-ray photoelectron spectroscopy (XPS), whereas physical changes were characterized using scanning electron microscopy (SEM) and optical imaging. Where applicable, photophysical characterization using extinction and fluorescence measurements were conducted. We demonstrate selective immobilization of QDs, and color-tunability using mixtures of QDs. QD-modified paper substrates were evaluated for photobleaching behavior, long-term storage, energy transfer improvements, model protease activity, and patterning via microcontact printing. The Au and Pt NP immobilized paper substrates were evaluated for SERS activity, and as supported catalysts in a model decolorization assay.  179  8.2 Results and discussion  8.2.1 Immobilization chemistries  Multiple chemistries were developed for the affinity immobilization of QDs and Au NPs on cellulose (Figure 8.1). The cellulose fibers of a paper substrate were either grafted with imidazole or dithiol(ane) groups or modified with dithiol(ane) groups after initial aminosilanization. For grafting, cellulose (1) was first oxidized to yield aldehyde groups (2) that were then coupled with imidazole (3) or dithiolane (4a) groups via reductive amination. For silanization, cellulose was first modified with (3-aminopropyl)triethoxysilane (APTES) (5), then reacted with a succinimidyl ester derivative of a dithiolane (6a). Dithiolane groups were reduced to dithiol groups (4b or 6b) as needed. The modified celluloses were characterized using a combination of infrared (IR) absorption spectroscopy, colorimetric chemical tests, and XPS.     Figure 8.1 Summary of the chemical modifications of cellulose (1): oxidation to yield aldehydes (2) enabled modification with imidazole groups (3) or dithiolane groups (4a). The latter were reduced to dithiol groups (4b). Cellulose (1) was also modified with APTES (5) and, subsequently, dithiolane (or dithiol) groups (6a and 6b). The R group may be another modification, an unreacted aldehyde (or the corresponding hydrate), or may have cyclized with the secondary amine of the shown modification. The chemistries were designed by Eleonora Petryayeva, who also contributed the XPS, IR, and colourimetric characterization data.  180  The IR data was diagnostic for the aminosilanization of (1) to (5), showing absorbance peaks corresponding to N-H bending at 1530–1600 cm–1. However, the IR was not capable of distinguishing between (5) and (6a/b). In contrast, XPS analysis of (5) and (6a) both showed peaks corresponding to N 1s from the aminosilanization, and S 2s and S 2p signals were prevalent in (6a). A colorimetric ninhydrin assay was employed to measure the amine groups in (5) and was estimated as 1.8–4.1 nmol mg–1 substrate. The number of thiols in (6b) was measured using the colorimetric Ellman assay and was estimated as 1.5 ± 0.2 nmol mg–1 substrate.  For modified cellulose substrates (2, 3, 4a), IR did not show any diagnostic peaks for relevant functional groups, which included aldehyde, amine, or thiol. Therefore, the number aldehyde groups after oxidation of (1) to (2) were measured via colorimetric assay with a 2,3,5-triphenyltetrazolium chloride (TTC)315,316, and was 23 ± 3 nmol mg–1 cellulose substrate. Similarly, the number of thiol groups for the conversion of (2) to (4b) was estimated to be 16 ± 3 nmol mg–1 substrate using a test with Ellman’s reagent.317 XPS analysis of these substrates showed weak N 1s peaks for (3) and (4a), which was negligible for (2). The dithiolane modified (4a) also showed relevant S 2s and S 2p peaks in the XPS data, suggesting effective modification with the ligand.  8.2.2 Immobilization of QDs  Aqueous colloidal QDs, coated with a variety of ligands, were immobilized on cellulose substrates (3), (4b), and (6b). These ligands included native hydrophobic ligands from the synthesis (e.g. TOPO), in addition to hydrophilic glutathione (GSH), 3-mercaptopropionic acid (MPA), cysteine, dihydrolipoic acid (DHLA), a sulfobetaine derivative of DHLA, and histidine (data not shown). The morphological differences between paper substrates with immobilized DHLA-QD630 were assessed using confocal 2PE and SEM (Figure 8.2A-B). The QD density scaled as (3) ≤ (4b) < (6b) and resulted in clearly visible fibers for (3), a matted appearance for (4b), and mostly indiscernible fibers for (6b). When kept hydrated, functionalization of the cellulose substrates with the small molecule ligands was necessary for the efficient immobilization of aqueous QDs. Comparisons between unmodified cellulose substrates and their modified counterparts (e.g. 3, 4a/b, 6a/b) demonstrate much brighter PL with the latter samples (Figure 8.2C). Interestingly, the QD PL did not scale with the apparent QD density on the paper substrates. The PL for QDs 181  immobilized on the (3) cellulose substrate were ~2-fold higher in comparison to the other substrates investigated. This is not uncommon as the thiol ligands within the 4b, and 6b substrates are known to quench QD PL significantly by acting as hole-acceptors, while the amine-based ligands such as those found in (3) are known to enhance QD PL.136  Beyond single colors of QD, it was possible to immobilize mixtures of red-emitting (QD640), green-emitting (QD525), and blue-emitting (QD450) QDs (all GSH-coated) on substrate (3) to produce net PL that spanned the rainbow in color, plus approximate white emission (Figure 8.2D). These emission colors outlined the expected red-green-blue (RGB) color triangle (Figure 8.2E) on a CIE 1931 diagram.   Figure 8.2 QDs immobilized on cellulose substrates (3a), (4b), and (6b). (A) 2PE PL images and (B) SEM images (secondary electron mode) at three different magnifications. The insets in the SEM images show the paper substrate without immobilized QDs and are at the same scale as the main images (and thus show a smaller region of interest). (C) PL spectra of modified cellulose substrates after exposure to aqueous QDs and washing. Errors bars are the standard deviation of at least three replicates and are shown only at the peak PL wavelength. Immobilization of multiple colors of QD. (D) Color photographs (smartphone camera) of selected mixtures of red-, green-, and blue-emitting QDs immobilized on circular (3 mm diameter) paper substrates (3). A blank substrate is also shown and has some background blue intensity. The numbers are sample identifiers. (E) Plot of the color of net PL emission for the samples from panel D (open circles) on the CIE 1931 color diagram. Data contributed by Eleonora Petryayeva and Hyungki Kim. 182  8.2.3 Patterning QDs by microcontact printing Some prospective applications of immobilized QDs, such as array-based sensing318 and anti-counterfeiting,319 benefit from patterning. The use of microcontact printing320 to pattern immobilized QDs on the modified cellulose substrates was therefore investigated. Polydimethylsiloxane (PDMS) stamps were prepared from 3D-printed templates because the methodology was a good match to the cellulose paper in terms of simplicity, low cost, and broad accessibility.  Dot and line patterns of hydrophobic alkyl-coated QDs were successfully printed on (3), including overlaid line patterns of green- and orange-emitting QDs (Figure 8.3A). Modification of the cellulose was necessary to retain the pattern of QDs after washing the substrate with solvent (Figure 8.3B). The printing of alkyl-coated QDs most easily yielded good pattern fidelity, as evaluated from the visibility of the line or dot features, the sharpness of their edges, and the uniformity of QD PL.      Figure 8.3. Patterned immobilization of QDs by microcontact printing. (A) Dot pattern of TOPO-QD605 (left) and crossed-lined patterns of TOPO-QD525 and TOPO-QD605 (right). Images were acquired without washing. (B) Comparison of patterning TOPO-QD605 on substrates (1) and (3), with and without washing post-stamping. Modification of the cellulose substrate was necessary for retention of the pattern.   Figure 8.4 compares the feature sizes of the 3D-printed mold, PDMS stamp, and line features of TOPO-QDs transferred by the PDMS stamp. The mold and the stamp had similar feature sizes (~300 μm) and the stamped QD features were ~10% larger (~330 μm). The slightly larger printed 183  feature size was expected from the elastomeric nature of PDMS and its deformation with applied pressure. Figure 8.5 shows an array of dots, where the mold had dot features with diameters of ~300 μm. The printed QD dot features were again ~10% larger.   Figure 8.4 Preparing line patterns. Optical image and example of a line profile plot for the 3-D printed mold (left), PDMS stamp (middle), and resulting alkyl-QD605 line pattern on (3) (right).   Figure 8.5 Image (left) and example of a line profile plot (right) for a dot array of alkyl-QD605.  184  Recognizing that hydrophobic QDs are less useful for prospective applications in bioanalysis, it was determined that at least partial exchange of TOPO ligands with GSH ligands was possible on-paper (3), post-immobilization of QDs. Figure 8.6 shows that Ellman’s assays indicated significantly higher numbers of thiols post-exposure to GSH for (3) with immobilized QD605 than (3) without. No thiols were detected for (3) alone or (3) with immobilized TOPO-QD605 (data not shown). These results suggested successful exchange. Unfortunately, at the time of the experiments, we were unable to obtain further confirmation by XPS.     Figure 8.6 GSH concentrations (determined by Ellman’s assay) for substrate (3), with and without immobilized TOPO-QD605, after exposure to a solution of GSH. The data shown was background subtracted from the signals from the respective control samples (i.e. (3) + TOPO-QD, or (3) alone).   In contrast to alkyl-QDs, printing hydrophilic GSH-coated QDs with good fidelity was more challenging, which was attributed to wetting of (3) by the QD solution being faster than immobilization. To address this challenge, the use of QDs coated with histidine (His) ligands instead of GSH ligands was investigated, as was the use of modified cellulose substrates (5) and (6). The binding of His ligands to QDs is weaker than that of GSH ligands, and we hypothesized that this switch would speed immobilization and thereby increase pattern fidelity. We hypothesized that (5) and (6) would yield better pattern fidelity because of the greater hydrophobicity and lower 185  wettability of these substrates, as well as the probability of fast immobilization from attractive electrostatic interactions between QDs and aminium groups associated with APTES. Indeed, it was possible to print patterns of His-QDs and GSH-QDs on substrates (5) and (6), albeit that immobilization of GSH-QDs with good fidelity also required the addition of 5% v/v glycerol to the stamping solution. Figure 8.7 shows pre-wash line patterns stamped using His-coated QD605 and GSH-coated QD605. The patterns from His-QDs more uniform than those from GSH-QDs, having more distinctive edges and more uniform PL intensity. Images of post-wash line patterns of His-QDs on (5) and (6a) are shown in Figure 8.8. Analogous data is shown for GSH-QDs in  Figure 8.9, including the effect of adding glycerol to the QD stamping solution.      Figure 8.7 PL micrographs for line patterns (pre-wash) made by stamping GSH-QD605 (top row) and His-QD605 (bottom row) on paper substrates (3). The scale bars are 500 μm.  186   Figure 8.8 PL micrographs for line patterns (post-wash) made by stamping His-QD605 on paper substrates (1), (5), and (6a). The scale bars are 500 μm.     Figure 8.9 PL micrographs for line patterns (post-wash) made by stamping GSH-QD605 on paper substrates (1), (5), and (6a), with and without glycerol in the stamping solution. The scale bars are 500 μm.   8.2.4 Immobilization of Au and Pt NPs In addition to QDs, metal NPs are other materials for which thiol and imidazole groups have affinity. The immobilization of aqueous, citrate-coated colloidal Au NPs (50 nm diameter) on (3), (4a), and (6a) was therefore tested (Figure 8.10). Pre-reduction of (4a/6a) to (4b/6b) was not 187  necessary because of the ability of Au NPs to spontaneously reduce disulfides. The visible color of the substrates, their extinction spectra, and SEM images all indicated that the density of Au NPs on the substrates scaled in the order (3) < (4a) < (6a). Here, density refers to both the total number of Au NPs per unit area and the number of clusters of many Au NPs. Plasmonic coupling between Au NPs in clusters was evident from the macroscopic blue/purple color from the immobilized NPs and the corresponding extinction spectra. For the latter, the plasmon absorption band for isolated Au NPs (~528 nm) was pronounced and did not tail to longer wavelengths with (3). The plasmon band was present but had a shoulder at longer wavelengths with (4a) and was present but was nearly overwhelmed by tailing at longer wavelengths with (6a). This trend was consistent with the SEM images.    Figure 8.10 Au NPs immobilized on substrates (3a), (4b), and (6b). (A) Photo-micrographs of the substrates without Au NPs (bottom/left) and with Au NPs (top/right). Images contributed by Hyungki Kim. (B) SEM images at three different length scales. The insets show the paper substrate without immobilized Au NPs and are at the same scale as the main images (and thus show a smaller region of interest). Note that the highest-magnification image for (4a) shows two regions of interest with Au NPs. SEM images for all substrates without Au NPs and for (6a) + Au NPs were acquired in secondary electron mode. SEM images for (3) and (4a) with Au NPs were acquired in back-scattered electron mode. (C) Photographs of paper substrates (diameter ~3 mm) with immobilized Au NPs (insets) and corresponding extinction spectra. Note that the extinction spectra are not corrected for the scattering contribution from the paper substrate. Spectra and images contributed by Hyungki Kim. 188  In contrast to the aqueous QDs, unmodified (1) or partially modified cellulose substrates (2) and (5) had some ability to retain Au NPs and Pt NPs, even when kept hydrated and washed. This difference was attributed to the broader ability of functional groups to coordinate to a noble metal surface than a semiconductor surface. Although significant, the density of immobilized Au NPs on (1) and (2) was less than (4a); however, (5) immobilized Au NPs at a density higher than (6a), which is attributed to strong electrostatic interactions between the anionic Au NPs and the aminium groups of (5). Another key difference between the immobilization of QDs and Au NPs was that the micromolar concentrations of QDs that saturated the substrates were readily obtained, whereas we were unable to achieve metal NP concentrations high enough to saturate the substrates.   8.2.5 Applications with immobilized metal NPs A potential application with immobilized metal NPs is as supported catalysts.314,321 As proof-of-concept, an assay for the reductive decolorization of methyl orange dye was investigated with (4a) paper substrates that were immobilized with Pt NPs or Au NPs (Figure 8.11A-B). The catalytic supports consisting of Pt NPs or Au NPs resulted in 100% methyl orange decolorization in the solution in comparison without the immobilized NPs. The initial-rates of decolorization were also faster, with a 2.2-fold, and 3.0-fold increase in the reaction rate relative to control substrates (i.e. without NPs). Immobilized metal NPs are also commonly utilized as substrates for SERS.322 Modified paper substrates (3, 4a, 6a) were soaked in solutions of 1,2-di(4-pyridyl)ethylene (BPE; a common SERS reporter 308,323) with or without prior immobilization with Au NPs (Figure 8.11C-D). The presence of Au NPs enhanced the SERS signal <2.4-fold with (3), ca. 3- to 5-fold with (4a), and ca. 7- to 11-fold with (6a).   189    Figure 8.11 (A) Putative mechanism for the Pt NP-catalyzed decolorization of methyl orange (MO) by sodium borohydride. (B) Methyl orange decolorization kinetics for (4b) with and without Pt NPs. The absorbance was measured at 466 nm and normalized to an initial value of unity. (C) Illustration showing the experimental format for SERS measurements. (D) Superposition of Raman spectra ob